*
* explain.cpp
* Explain query execution plans
*
* Portions Copyright (c) 2020 Huawei Technologies Co.,Ltd.
* Portions Copyright (c) 1996-2012, PostgreSQL Global Development Group
* Portions Copyright (c) 1994-5, Regents of the University of California
*
* IDENTIFICATION
* src/gausskernel/optimizer/commands/explain.cpp
*
* -------------------------------------------------------------------------
*/
#include "postgres.h"
#include "knl/knl_variable.h"
#include "access/tableam.h"
#include "access/xact.h"
#include "catalog/indexing.h"
#include "catalog/namespace.h"
#include "catalog/pg_collation.h"
#include "catalog/gs_obsscaninfo.h"
#include "catalog/pg_obsscaninfo.h"
#include "catalog/pg_type.h"
#include "db4ai/create_model.h"
#include "db4ai/hyperparameter_validation.h"
#include "commands/createas.h"
#include "commands/defrem.h"
#include "commands/prepare.h"
#include "executor/node/nodeModifyTable.h"
#include "executor/exec/execStream.h"
#include "executor/hashjoin.h"
#include "executor/lightProxy.h"
#include "executor/node/nodeAgg.h"
#include "executor/node/nodeRecursiveunion.h"
#include "executor/node/nodeSetOp.h"
#include "foreign/dummyserver.h"
#include "foreign/fdwapi.h"
#include "instruments/generate_report.h"
#include "nodes/print.h"
#include "opfusion/opfusion_util.h"
#include "opfusion/opfusion.h"
#include "optimizer/cost.h"
#include "optimizer/clauses.h"
#include "optimizer/dataskew.h"
#include "optimizer/dynsmp.h"
#include "optimizer/ml_model.h"
#include "optimizer/nodegroups.h"
#include "optimizer/streamplan.h"
#include "optimizer/planmem_walker.h"
#include "optimizer/planner.h"
#include "optimizer/randomplan.h"
#include "parser/parse_hint.h"
#include "parser/parsetree.h"
#include "rewrite/rewriteHandler.h"
#include "tcop/tcopprot.h"
#include "utils/builtins.h"
#include "utils/hotkey.h"
#include "utils/json.h"
#include "utils/lsyscache.h"
#include "utils/rel.h"
#include "utils/rel_gs.h"
#include "utils/snapmgr.h"
#include "utils/tuplesort.h"
#include "utils/typcache.h"
#include "utils/xml.h"
#include "utils/batchsort.h"
#include "vecexecutor/vechashagg.h"
#include "vecexecutor/vechashjoin.h"
#include "vecexecutor/vecsetop.h"
#include "vectorsonic/vsonichashagg.h"
#include "workload/cpwlm.h"
#include "workload/workload.h"
#include "pgxc/execRemote.h"
#ifdef PGXC
#include "pgxc/groupmgr.h"
#include "catalog/pgxc_node.h"
#include "pgxc/pgxc.h"
#endif
#include "db4ai/aifuncs.h"
THR_LOCAL bool OnlySelectFromPlanTable = false;
THR_LOCAL bool OnlyDeleteFromPlanTable = false;
THR_LOCAL bool PTFastQueryShippingStore = true;
THR_LOCAL bool IsExplainPlanStmt = false;
THR_LOCAL bool IsExplainPlanSelectForUpdateStmt = false;
THR_LOCAL explain_get_index_name_hook_type explain_get_index_name_hook = NULL;
extern TrackDesc trackdesc[];
extern sortMessage sortmessage[];
extern DestReceiver* CreateDestReceiver(CommandDest dest);
static const PlanTableEntry plan_table_cols[] = {{"id", PLANID, INT4OID, ANAL_OPT},
{"operation", PLAN, TEXTOID, ANAL_OPT},
{"A-time", ACTUAL_TIME, TEXTOID, ANAL_OPT},
{"P-time", PREDICT_TIME, TEXTOID, PRED_OPT_TIME},
{"A-rows", ACTUAL_ROWS, FLOAT8OID, ANAL_OPT},
{"P-rows", PREDICT_ROWS, FLOAT8OID, PRED_OPT_ROW},
{"E-rows", ESTIMATE_ROWS, FLOAT8OID, COST_OPT},
{"E-distinct", ESTIMATE_DISTINCT, TEXTOID, VERBOSE_OPT},
{"Peak Memory", ACTUAL_MEMORY, TEXTOID, ANAL_OPT},
{"P-memory", PREDICT_MEMORY, TEXTOID, PRED_OPT_MEM},
{"E-memory", ESTIMATE_MEMORY, TEXTOID, COST_OPT},
{"A-width", ACTUAL_WIDTH, TEXTOID, ANAL_OPT},
{"E-width", ESTIMATE_WIDTH, INT4OID, COST_OPT},
{"E-costs", ESTIMATE_COSTS, TEXTOID, COST_OPT}
};
#define NUM_VER_COLS 1
#define NUM_ANALYZE_COLS 4
#define NUM_COST_COLS 4
#define MIN_FILE_NUM 1024
#define MIN_DISK_USED 1024
#define X_OPENING 0
#define X_CLOSING 1
#define X_CLOSE_IMMEDIATE 2
#define X_NOWHITESPACE 4
#define PUSHDOWN_PREDICATE_FLAG "Pushdown Predicate Filter"
#define PUSHDOWN_FILTER_INFORMATIONAL_CONSTRAINT_FLAG "Pushdown Predicate Filter(Informational Constraint Optimization)"
#define FILTER_INFORMATIONAL_CONSTRAINT_FLAG "Filter(Informational Constraint Optimization)"
#define FILTER_FLAG "Filter"
static void ExplainOneQuery(
Query* query, IntoClause* into, ExplainState* es, const char* queryString, DestReceiver *dest, ParamListInfo params);
static void report_triggers(ResultRelInfo* rInfo, bool show_relname, ExplainState* es);
template <bool is_pretty>
static void ExplainNode(
PlanState* planstate, List* ancestors, const char* relationship, const char* plan_name, ExplainState* es);
static void CalculateProcessedRows(
PlanState* planstate, int idx, int smpIdx, const uint64* innerRows, const uint64* outterRows, uint64* processedrows);
static void show_plan_execnodes(PlanState* planstate, ExplainState* es);
static void show_plan_tlist(PlanState* planstate, List* ancestors, ExplainState* es);
static bool ContainRlsQualInRteWalker(Node* node, void* unused_info);
static void show_expression(
Node* node, const char* qlabel, PlanState* planstate, List* ancestors, bool useprefix, ExplainState* es);
static void show_qual(
List* qual, const char* qlabel, PlanState* planstate, List* ancestors, bool useprefix, ExplainState* es);
static void show_scan_qual(List *qual, const char *qlabel, PlanState *planstate, List *ancestors, ExplainState *es,
bool show_prefix = false);
static void show_skew_optimization(const PlanState* planstate, ExplainState* es);
template <bool generate>
static void show_bloomfilter(Plan* plan, PlanState* planstate, List* ancestors, ExplainState* es);
template <bool generate>
static void show_bloomfilter_number(const List* filterNumList, ExplainState* es);
* @Description: Show pushdown quals in the flag identifier.
* @in planstate: Keep the current state of the plan node.
* @in ancestors: A list of parent PlanState nodes, most-closely-nested first.
* @in identifier: The label of pushdown predicate clause.
* @in es: A ExplainState struct.
* @return None.
*/
static void show_pushdown_qual(PlanState* planstate, List* ancestors, ExplainState* es, const char* identifier);
static void show_upper_qual(List* qual, const char* qlabel, PlanState* planstate, List* ancestors, ExplainState* es);
static void show_groupby_keys(AggState* aggstate, List* ancestors, ExplainState* es);
static void show_sort_keys(SortState* sortstate, List* ancestors, ExplainState* es);
static void show_merge_append_keys(MergeAppendState* mstate, List* ancestors, ExplainState* es);
static void show_merge_sort_keys(PlanState* state, List* ancestors, ExplainState* es);
static void show_startwith_pseudo_entries(PlanState* state, List* ancestors, ExplainState* es);
static void show_sort_info(SortState* sortstate, ExplainState* es);
static void show_sort_group_info(SortGroupState *state, ExplainState *es);
static void show_hash_info(HashState* hashstate, ExplainState* es);
static void show_vechash_info(VecHashJoinState* hashstate, ExplainState* es);
static void show_tidbitmap_info(BitmapHeapScanState *planstate, ExplainState *es);
static void show_instrumentation_count(const char* qlabel, int which, const PlanState* planstate, ExplainState* es);
static void show_removed_rows(int which, const PlanState* planstate, int idx, int smpIdx, int* removeRows);
static int check_integer_overflow(double var);
static void show_foreignscan_info(ForeignScanState* fsstate, ExplainState* es);
static void show_detail_storage_info_text(Instrumentation* instr, StringInfo instr_info);
static void show_detail_storage_info_json(Instrumentation* instr, StringInfo instr_info, ExplainState* es);
static void show_storage_filter_info(PlanState* planstate, ExplainState* es);
static void show_llvm_info(const PlanState* planstate, ExplainState* es);
static void show_modifytable_merge_info(const PlanState* planstate, ExplainState* es);
static void show_recursive_info(RecursiveUnionState* rustate, ExplainState* es);
static void show_startwith_dfx(StartWithOpState* rustate, ExplainState* es);
static const char* explain_get_index_name(Oid indexId);
static void ExplainIndexScanDetails(Oid indexid, ScanDirection indexorderdir, ExplainState* es);
static void ExplainScanTarget(Scan* plan, ExplainState* es);
static void ExplainModifyTarget(ModifyTable* plan, ExplainState* es);
static void ExplainTargetRel(Plan* plan, Index rti, ExplainState* es, bool multiTarget = false);
static void show_on_duplicate_info(ModifyTableState* mtstate, ExplainState* es, List* ancestors);
#ifndef PGXC
static void show_modifytable_info(ModifyTableState* mtstate, ExplainState* es);
#endif
static void ExplainMemberNodes(const List* plans, PlanState** planstates, List* ancestors, ExplainState* es);
static void ExplainSubPlans(List* plans, List* ancestors, const char* relationship, ExplainState* es);
static void ExplainProperty(const char* qlabel, const char* value, bool numeric, ExplainState* es);
static void ExplainOpenGroup(const char* objtype, const char* labelname, bool labeled, ExplainState* es);
static void ExplainCloseGroup(const char* objtype, const char* labelname, bool labeled, ExplainState* es);
static void ExplainDummyGroup(const char* objtype, const char* labelname, ExplainState* es);
#ifdef PGXC
static void ExplainExecNodes(const ExecNodes* en, ExplainState* es);
static void ExplainRemoteQuery(RemoteQuery* plan, PlanState* planstate, List* ancestors, ExplainState* es);
#endif
static void ExplainXMLTag(const char* tagname, unsigned int flags, ExplainState* es);
static void ExplainJSONLineEnding(ExplainState* es);
static void ExplainYAMLLineStarting(ExplainState* es);
static void escape_yaml(StringInfo buf, const char* str);
template <bool is_detail>
static void show_time(ExplainState* es, const Instrumentation* instrument, int idx);
static void show_cpu(ExplainState* es, const Instrumentation* instrument, double innerCycles, double outerCycles, int nodeIdx,
int smpIdx, uint64 proRows);
static void show_detail_cpu(ExplainState* es, PlanState* planstate);
static void show_datanode_buffers(ExplainState* es, PlanState* planstate);
static void show_buffers(ExplainState* es, StringInfo infostr, const Instrumentation* instrument, bool is_datanode,
int nodeIdx, int smpIdx, const char* nodename);
static void show_datanode_time(ExplainState* es, PlanState* planstate);
static void ShowStreamRunNodeInfo(Stream* stream, ExplainState* es);
static void ShowRunNodeInfo(const ExecNodes* en, ExplainState* es, const char* qlabel);
template <bool is_detail>
static void show_datanode_hash_info(ExplainState *es, int nbatch, int nbuckets_original, int nbatch_original,
int nbuckets, long spacePeakKb);
static void ShowRoughCheckInfo(ExplainState* es, Instrumentation* instrument, int nodeIdx, int smpIdx);
static void show_hashAgg_info(AggState* hashaggstate, ExplainState* es);
static void ExplainPrettyList(List* data, ExplainState* es);
static void show_pretty_time(ExplainState* es, Instrumentation* instrument, char* node_name, int nodeIdx, int smpIdx,
int dop, bool executed = true);
static void show_analyze_buffers(ExplainState* es, const PlanState* planstate, StringInfo infostr, int nodeNum);
inline static void show_cpu_info(StringInfo infostr, double incCycles, double exCycles, uint64 proRows);
static void show_track_time_info(ExplainState* es);
template <bool datanode>
static void show_child_cpu_cycles_and_rows(PlanState* planstate, int idx, int smpIdx, double* outerCycles,
double* innerCycles, uint64* outterRows, uint64* innerRows);
template <bool datanode>
static void CalCPUMemberNode(const List* plans, PlanState** planstates, int idx, int smpIdx, double* Cycles);
static void showStreamnetwork(Stream* stream, ExplainState* es);
static int get_digit(int value);
template <bool datanode>
static void get_oper_time(ExplainState* es, PlanState* planstate, const Instrumentation* instr, int nodeIdx, int smpIdx);
static void show_peak_memory(ExplainState* es, int plan_size);
static bool get_execute_mode(const ExplainState* es, int idx);
static void show_setop_info(SetOpState* setopstate, ExplainState* es);
static void show_grouping_sets(PlanState* planstate, Agg* agg, List* ancestors, ExplainState* es);
static void show_group_keys(GroupState* gstate, List* ancestors, ExplainState* es);
static void show_sort_group_keys(PlanState* planstate, const char* qlabel, int nkeys, const AttrNumber* keycols,
const Oid* sortOperators, const Oid* collations, const bool* nullsFirst, List* ancestors, ExplainState* es);
static void show_sortorder_options(StringInfo buf, const Node* sortexpr, Oid sortOperator, Oid collation, bool nullsFirst);
static void show_grouping_set_keys(PlanState* planstate, Agg* aggnode, Sort* sortnode, List* context, bool useprefix,
List* ancestors, ExplainState* es);
static void show_dn_executor_time(ExplainState* es, int plan_node_id, ExecutorTime stage);
static void show_stream_send_time(ExplainState* es, const PlanState* planstate);
static void append_datanode_name(ExplainState* es, char* node_name, int dop, int j);
static void show_tablesample(Plan* plan, PlanState* planstate, List* ancestors, ExplainState* es);
void insert_obsscaninfo(
uint64 queryid, const char* rel_name, int64 file_count, double scan_data_size, double total_time, int format);
extern List* get_str_targetlist(List* fdw_private);
static void show_wlm_explain_name(ExplainState* es, const char* plan_name, const char* pname, int plan_node_id);
extern unsigned char pg_toupper(unsigned char ch);
static char* set_strtoupper(const char* str, uint32 len);
#ifdef ENABLE_MULTIPLE_NODES
static bool show_scan_distributekey(const Plan* plan)
{
return (
IsA(plan, CStoreScan) || IsA(plan, CStoreIndexScan) || IsA(plan, CStoreIndexHeapScan) || IsA(plan, SeqScan) ||
IsA(plan, IndexScan) || IsA(plan, IndexOnlyScan) || IsA(plan, CteScan) || IsA(plan, AnnIndexScan) ||
IsA(plan, ForeignScan) || IsA(plan, VecForeignScan) || IsA(plan, BitmapHeapScan) || IsA(plan, TsStoreScan)
);
}
#endif
static void show_unique_check_info(PlanState *planstate, ExplainState *es);
static void show_ndpplugin_statistic(ExplainState *es, PlanState* planstate, StringInfo str, bool is_pretty);
* ExplainQuery -
* execute an EXPLAIN command
*/
void ExplainQuery(
ExplainStmt* stmt, const char* queryString, ParamListInfo params, DestReceiver* dest, char* completionTag)
{
ExplainState es;
TupOutputState* tstate = NULL;
List* rewritten = NIL;
ListCell* lc = NULL;
bool timing_set = false;
ExplainInitState(&es);
foreach (lc, stmt->options) {
DefElem* opt = (DefElem*)lfirst(lc);
* For explain plan stmt:
* 1.Cache the plan instead of printing it, and return 'EXPLAIN SUCCESS'.
* 2.Insert the cached info into plan_table, then user can get plan info when select from table;
* 3.As the explain option "PLAN" cannot work with other options,
* when other options are detected, report as an error.
* 4.And we do not support execute it on DN now.
*
*/
if (strcmp(opt->defname, "plan") == 0) {
es.plan = defGetBoolean(opt);
if (es.plan) {
#ifdef ENABLE_MULTIPLE_NODES
if (IS_PGXC_DATANODE || IS_SINGLE_NODE)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("EXPLAIN PLAN does not support on datanode or single node.")));
#endif
if (list_length(stmt->options) != 1)
ereport(ERROR,
(errcode(ERRCODE_SYNTAX_ERROR),
errmsg("EXPLAIN option 'PLAN' can not work with other options.")));
IsExplainPlanStmt = true;
* To collect plan info for explain plan, we don't use FQS.
* However, select for update can only run when enable_fast_query_shipping = on,
* so we should set IsExplainPlanSelectForUpdateStmt = true to make enable_fast_query_shipping work.
*/
if (IsA(stmt->query, Query) && ((Query*)stmt->query)->hasForUpdate)
IsExplainPlanSelectForUpdateStmt = true;
if (PTFastQueryShippingStore == true && IsExplainPlanSelectForUpdateStmt == false)
PTFastQueryShippingStore = false;
if (stmt->statement == NULL)
es.statement_id = NULL;
else {
Value v = ((A_Const*)stmt->statement)->val;
if (v.type == T_Null)
es.statement_id = NULL;
else
es.statement_id = v.val.str;
}
dest->mydest = DestNone;
t_thrd.explain_cxt.explain_perf_mode = EXPLAIN_PRETTY;
errno_t errorno =
snprintf_s(completionTag, COMPLETION_TAG_BUFSIZE, COMPLETION_TAG_BUFSIZE - 1, "EXPLAIN SUCCESS");
securec_check_ss(errorno, "\0", "\0");
}
} else if (strcmp(opt->defname, "analyze") == 0)
es.analyze = defGetBoolean(opt);
else if (strcmp(opt->defname, "verbose") == 0)
es.verbose = defGetBoolean(opt);
else if (strcmp(opt->defname, "costs") == 0)
es.costs = defGetBoolean(opt);
else if (strcmp(opt->defname, "buffers") == 0)
es.buffers = defGetBoolean(opt);
#if defined(ENABLE_MULTIPLE_NODES) || defined(USE_SPQ)
else if (strcmp(opt->defname, "nodes") == 0 && IS_SPQ_RUNNING)
es.nodes = defGetBoolean(opt);
else if (strcmp(opt->defname, "num_nodes") == 0 && IS_SPQ_RUNNING)
es.num_nodes = defGetBoolean(opt);
else if (pg_strcasecmp(opt->defname, "detail") == 0 && IS_SPQ_RUNNING)
es.detail = defGetBoolean(opt);
#endif
else if (strcmp(opt->defname, "timing") == 0) {
timing_set = true;
es.timing = defGetBoolean(opt);
} else if (pg_strcasecmp(opt->defname, "cpu") == 0)
es.cpu = defGetBoolean(opt);
else if (pg_strcasecmp(opt->defname, "performance") == 0)
es.performance = defGetBoolean(opt);
else if (strcmp(opt->defname, "format") == 0) {
char* p = defGetString(opt);
if (strcmp(p, "text") == 0)
es.format = EXPLAIN_FORMAT_TEXT;
else if (strcmp(p, "xml") == 0)
es.format = EXPLAIN_FORMAT_XML;
else if (strcmp(p, "json") == 0)
es.format = EXPLAIN_FORMAT_JSON;
else if (strcmp(p, "yaml") == 0)
es.format = EXPLAIN_FORMAT_YAML;
else
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("unrecognized value for EXPLAIN option \"%s\": \"%s\"", opt->defname, p)));
#ifndef ENABLE_MULTIPLE_NODES
} else if (strcmp(opt->defname, "predictor") == 0) {
es.opt_model_name = defGetString(opt);
#endif
} else
ereport(ERROR, (errcode(ERRCODE_SYNTAX_ERROR), errmsg("unrecognized EXPLAIN option \"%s\"", opt->defname)));
}
if (es.performance) {
es.analyze = true;
es.buffers = true;
es.costs = true;
es.cpu = true;
es.detail = true;
#ifdef ENABLE_MULTIPLE_NODES
#ifdef PGXC
es.nodes = true;
es.num_nodes = true;
#endif
#endif
es.timing = true;
es.verbose = true;
}
if (es.buffers && !es.analyze)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("EXPLAIN option BUFFERS requires ANALYZE")));
if (es.cpu && !es.analyze)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("EXPLAIN option CPU requires ANALYZE")));
if (es.detail && !es.analyze)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("EXPLAIN option DETAIL requires ANALYZE")));
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es.format != EXPLAIN_FORMAT_TEXT)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("explain_perf_mode requires FORMAT TEXT")));
es.timing = (timing_set) ? es.timing : es.analyze;
if (es.timing && !es.analyze)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("EXPLAIN option TIMING requires ANALYZE")));
* Parse analysis was done already, but we still have to run the rule
* rewriter. We do not do AcquireRewriteLocks: we assume the query either
* came straight from the parser, or suitable locks were acquired by
* plancache.c.
*/
AssertEreport(IsA(stmt->query, Query), MOD_EXECUTOR, "unexpect query type");
rewritten = QueryRewrite(castNode(Query, stmt->query));
ExplainBeginOutput(&es);
if (rewritten == NIL) {
* In the case of an INSTEAD NOTHING, tell at least that. But in
* non-text format, the output is delimited, so this isn't necessary.
*/
if (es.format == EXPLAIN_FORMAT_TEXT)
appendStringInfoString(es.str, "Query rewrites to nothing\n");
* In centralized mode, non-stream plans only support EXPLAIN_NORMAL.
* So set explain_perf_mode to EXPLAIN_NORMAL here.
*/
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL &&
!(IS_STREAM_PLAN && u_sess->exec_cxt.under_stream_runtime)) {
t_thrd.explain_cxt.explain_perf_mode = EXPLAIN_NORMAL;
}
} else {
ListCell* l = NULL;
t_thrd.explain_cxt.explain_light_proxy = IS_PGXC_COORDINATOR && !IsConnFromCoord() &&
u_sess->attr.attr_sql.enable_light_proxy &&
(list_length(rewritten) == 1) && !es.analyze;
if (u_sess->attr.attr_sql.enable_autoanalyze && es.analyze &&
t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
MemoryContext oldcontext = MemoryContextSwitchTo(u_sess->temp_mem_cxt);
u_sess->analyze_cxt.autoanalyze_timeinfo = makeStringInfo();
(void)MemoryContextSwitchTo(oldcontext);
} else
u_sess->analyze_cxt.autoanalyze_timeinfo = NULL;
if (IS_PGXC_COORDINATOR && !IsConnFromCoord() && es.analyze) {
t_thrd.postgres_cxt.mark_explain_analyze = true;
t_thrd.postgres_cxt.mark_explain_only = false;
} else {
t_thrd.postgres_cxt.mark_explain_analyze = false;
t_thrd.postgres_cxt.mark_explain_only = true;
}
foreach (l, rewritten) {
Query* query_tree = (Query*)lfirst(l);
* We need to revert this query_tree, so that ExplainOneQuery can get correct
* message when generating PlanedStmt.
*/
query_tree->is_flt_frame = !query_check_no_flt(query_tree);
ExplainOneQuery(query_tree, NULL, &es, queryString, None_Receiver, params);
if (lnext(l) != NULL)
ExplainSeparatePlans(&es);
}
t_thrd.explain_cxt.explain_light_proxy = false;
}
ExplainEndOutput(&es);
AssertEreport(es.indent == 0, MOD_EXECUTOR, "unexpect es.indent value");
if (!es.plan) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
es.planinfo->dump(dest);
} else {
tstate = begin_tup_output_tupdesc(dest, ExplainResultDesc(stmt));
if (es.format == EXPLAIN_FORMAT_TEXT)
(void)do_text_output_multiline(tstate, es.str->data);
else
do_text_output_oneline(tstate, es.str->data);
end_tup_output(tstate);
}
}
#if defined(ENABLE_MULTIPLE_NODES) || defined(USE_SPQ)
if (IS_SPQ_RUNNING && u_sess->instr_cxt.global_instr != NULL) {
delete u_sess->instr_cxt.global_instr;
u_sess->instr_cxt.thread_instr = NULL;
u_sess->instr_cxt.global_instr = NULL;
}
#endif
if (u_sess->analyze_cxt.autoanalyze_timeinfo != NULL) {
pfree_ext(u_sess->analyze_cxt.autoanalyze_timeinfo->data);
pfree_ext(u_sess->analyze_cxt.autoanalyze_timeinfo);
u_sess->analyze_cxt.autoanalyze_timeinfo = NULL;
}
stmt->planinfo = es.planinfo;
pfree_ext(es.str->data);
list_free_deep(rewritten);
}
* Initialize ExplainState.
*/
void ExplainInitState(ExplainState* es)
{
errno_t rc = memset_s(es, sizeof(ExplainState), 0, sizeof(ExplainState));
securec_check(rc, "\0", "\0");
es->costs = true;
#ifdef ENABLE_MULTIPLE_NODES
#ifdef PGXC
es->nodes = true;
#endif
#endif
es->str = makeStringInfo();
es->planinfo = NULL;
es->from_dn = false;
es->sql_execute = true;
es->isexplain_execute = false;
es->datanodeinfo.all_datanodes = true;
es->datanodeinfo.len_nodelist = 0;
es->datanodeinfo.node_index = NULL;
es->indent = 0;
es->pindent = 0;
es->wlm_statistics_plan_max_digit = NULL;
es->es_frs.parent = es;
IsExplainPlanStmt = false;
IsExplainPlanSelectForUpdateStmt = false;
}
* ExplainResultDesc -
* construct the result tupledesc for an EXPLAIN
*/
TupleDesc ExplainResultDesc(ExplainStmt* stmt)
{
TupleDesc tupdesc = NULL;
ListCell* lc = NULL;
Oid result_type = TEXTOID;
char* attributeName = NULL;
char RandomPlanName[NAMEDATALEN] = {0};
bool explain_plan = false;
foreach (lc, stmt->options) {
DefElem* opt = (DefElem*)lfirst(lc);
if (strcmp(opt->defname, "format") == 0) {
char* p = defGetString(opt);
if (strcmp(p, "xml") == 0)
result_type = XMLOID;
else if (strcmp(p, "json") == 0)
result_type = JSONOID;
else
result_type = TEXTOID;
} else if (strcmp(opt->defname, "plan") == 0 && defGetBoolean(opt) && list_length(stmt->options) == 1) {
explain_plan = true;
}
}
if (!explain_plan) {
tupdesc = CreateTemplateTupleDesc(1, false);
if (u_sess->attr.attr_sql.plan_mode_seed != OPTIMIZE_PLAN) {
int rc = 0;
rc = snprintf_s(
RandomPlanName, NAMEDATALEN, NAMEDATALEN - 1, "QUERY PLAN (RANDOM seed %u)", get_inital_plan_seed());
securec_check_ss(rc, "\0", "\0");
}
attributeName =
((OPTIMIZE_PLAN == u_sess->attr.attr_sql.plan_mode_seed) ? (char*)"QUERY PLAN" : RandomPlanName);
TupleDescInitEntry(tupdesc, (AttrNumber)1, attributeName, result_type, -1, 0);
}
return tupdesc;
}
* ExplainOneQuery -
* print out the execution plan for one Query
*
* "into" is NULL unless we are explaining the contents of a CreateTableAsStmt.
*/
static void ExplainOneQuery(
Query* query, IntoClause* into, ExplainState* es, const char* queryString, DestReceiver *dest, ParamListInfo params)
{
if (t_thrd.explain_cxt.explain_light_proxy) {
ExecNodes* single_exec_node = lightProxy::checkLightQuery(query);
if (single_exec_node == NULL || list_length(single_exec_node->nodeList) +
list_length(single_exec_node->primarynodelist) != 1) {
t_thrd.explain_cxt.explain_light_proxy = false;
}
CleanHotkeyCandidates(true);
}
u_sess->exec_cxt.remotequery_list = NIL;
if (query->commandType == CMD_UTILITY) {
if (IsA(query->utilityStmt, CreateTableAsStmt)) {
CreateTableAsStmt* ctas = (CreateTableAsStmt*)query->utilityStmt;
List* rewritten = NIL;
if (!es->analyze) {
const char* stmt = ctas->is_select_into ? "SELECT INTO" : "CREATE TABLE AS SELECT";
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("EXPLAIN %s requires ANALYZE", stmt)));
}
rewritten = QueryRewriteCTAS(query);
ExplainOneQuery((Query*)linitial(rewritten), ctas->into, es, queryString, dest, params);
return;
}
else if (IsA(query->utilityStmt, CreateModelStmt)) {
CreateModelStmt* cm = (CreateModelStmt*) query->utilityStmt;
* Create the tuple receiver object and insert hyperp it will need
*/
DestReceiverTrainModel* dest_train_model = NULL;
dest_train_model = (DestReceiverTrainModel*) CreateDestReceiver(DestTrainModel);
configure_dest_receiver_train_model(dest_train_model, CurrentMemoryContext, (AlgorithmML)cm->algorithm,
cm->model, queryString, true);
PlannedStmt *plan =
plan_create_model(cm, queryString, params, (DestReceiver *)dest_train_model, CurrentMemoryContext);
print_hyperparameters(DEBUG1, dest_train_model->hyperparameters);
ExplainOnePlan(plan, into, es, queryString, dest, params);
return;
}
ExplainOneUtility(query->utilityStmt, into, es, queryString, params);
return;
}
* Query is not actually executed when declaring a cursor, but will generate a plan.
* So we can use explain to print the plan, but can not use explain analyze to execute the query.
*/
if (query->utilityStmt && IsA(query->utilityStmt, DeclareCursorStmt)) {
const char* s = es->performance ? "PERFORMANCE" : "ANALYZE";
if (es->analyze)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("EXPLAIN %s is not supported when declaring a cursor.", s),
errdetail("Query is not actually executed when declaring a cursor.")));
}
PlannedStmt* plan = NULL;
* plan the query
* Temporarily apply SET hint using PG_TRY for later recovery
*/
int nest_level = apply_set_hint(query);
AFTER_EXPLAIN_APPLY_SET_HINT();
PG_TRY();
{
plan = pg_plan_query(query, CURSOR_OPT_SPQ_OK, params, true);
}
PG_CATCH();
{
recover_set_hint(nest_level);
AFTER_EXPLAIN_RECOVER_SET_HINT();
PG_RE_THROW();
}
PG_END_TRY();
recover_set_hint(nest_level);
AFTER_EXPLAIN_RECOVER_SET_HINT();
CleanHotkeyCandidates(true);
check_gtm_free_plan((PlannedStmt *)plan, (es->analyze && !ClusterResizingInProgress()) ? ERROR : WARNING);
check_plan_mergeinto_replicate(plan, es->analyze ? ERROR : WARNING);
es->is_explain_gplan = true;
ExplainOnePlan(plan, into, es, queryString, dest, params);
}
* ExplainOneUtility -
* print out the execution plan for one utility statement
* (In general, utility statements don't have plans, but there are some
* we treat as special cases)
*
* "into" is NULL unless we are explaining the contents of a CreateTableAsStmt.
*
* This is exported because it's called back from prepare.c in the
* EXPLAIN EXECUTE case.
*/
void ExplainOneUtility(
Node* utilityStmt, IntoClause* into, ExplainState* es, const char* queryString, ParamListInfo params)
{
if (utilityStmt == NULL)
return;
if (IsA(utilityStmt, ExecuteStmt))
ExplainExecuteQuery((ExecuteStmt*)utilityStmt, into, es, queryString, params);
else if (IsA(utilityStmt, NotifyStmt)) {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoString(es->str, "NOTIFY\n");
else
ExplainDummyGroup("Notify", NULL, es);
} else {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoString(es->str, "Utility statements have no plan structure\n");
else
ExplainDummyGroup("Utility Statement", NULL, es);
}
}
* @Description: read data from channel and get data before print performance data
* @in estate - estate information
* @out - void
*/
static void ExecRemoteprocessPlan(EState* estate)
{
#ifdef ENABLE_MULTIPLE_NODES
ListCell* lc = NULL;
foreach (lc, estate->es_remotequerystates) {
PlanState* ps = (PlanState*)lfirst(lc);
ExecEndRemoteQuery((RemoteQueryState*)ps, true);
}
#else
if (u_sess->stream_cxt.global_obj) {
u_sess->stream_cxt.global_obj->SigStreamThreadClose();
StreamNodeGroup::syncQuit(STREAM_COMPLETE);
}
#endif
}
* @Description: Find real exec dn nodes within all RemoteQuery PlanNodes.
* @in qd - queryDesc information
* @out - void
*/
static void GetNodeList(QueryDesc* qd)
{
ListCell* lc = NULL;
ListCell* lc2 = NULL;
RemoteQuery* rq = NULL;
foreach (lc, u_sess->exec_cxt.remotequery_list) {
rq = (RemoteQuery*)lfirst(lc);
* when execute pbe query, rq->exec_nodes->nodeList could be NIL,
* so we should get nodeLists through FindExecNodesInPBE function.
*/
foreach (lc2, qd->estate->es_remotequerystates) {
RemoteQueryState* rqs = (RemoteQueryState*)lfirst(lc2);
FindExecNodesInPBE(rqs, rq->exec_nodes, EXEC_ON_DATANODES);
}
}
}
* @Description: Reorganize the executed query which will be pushed down to DN node.
* @in es - explainstate information
* @in rq - remote query plannode
* @in queryString - prepared sql statement in pbe or simple queryString
* @in explain_sql - executed sql statement in pbe
* @in nodeoid - DN nodeoid
* @out - void
*/
void ReorganizeSqlStatement(
ExplainState* es, RemoteQuery* rq, const char* queryString, const char* explain_sql, Oid nodeoid)
{
const char* perf_mode = NULL;
switch (u_sess->attr.attr_sql.guc_explain_perf_mode) {
case EXPLAIN_NORMAL:
perf_mode = "normal";
break;
case EXPLAIN_PRETTY:
perf_mode = "pretty";
break;
case EXPLAIN_SUMMARY:
perf_mode = "summary";
break;
case EXPLAIN_RUN:
perf_mode = "run";
break;
default:
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("invalid value for parameter explain_perf_mode.")));
break;
}
const char* format = NULL;
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
format = "text";
break;
case EXPLAIN_FORMAT_XML:
format = "xml";
break;
case EXPLAIN_FORMAT_JSON:
format = "json";
break;
case EXPLAIN_FORMAT_YAML:
format = "yaml";
break;
default:
ereport(ERROR,
(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("unrecognized value for EXPLAIN option \"format\".")));
break;
}
const char* verbose = es->verbose ? "on" : "off";
const char* costs = es->costs ? "on" : "off";
const char* num_nodes = es->num_nodes ? "on" : "off";
StringInfo prefix = makeStringInfo();
appendStringInfo(prefix,
"set explain_perf_mode=%s; "
"explain (format %s, verbose %s, costs %s, nodes off, num_nodes %s)",
perf_mode,
format,
verbose,
costs,
num_nodes);
if (es->isexplain_execute && (!CheckPrepared(rq, nodeoid) || ENABLE_GPC)) {
* When execute firstly explain pbe statement in dfx, we should send
* prepared statement and execute statement at the same time.
* Also, we should not need to deallocate prepare statement, because
* prepared statement in DN will only be sent once.
*/
StringInfo sql = makeStringInfo();
appendStringInfo(sql, "%s; %s %s", queryString, prefix->data, explain_sql);
rq->execute_statement = sql->data;
} else {
StringInfo sql = makeStringInfo();
appendStringInfo(sql, "%s %s", prefix->data, explain_sql);
rq->execute_statement = sql->data;
}
}
* @Description: Execute FQS query in DN
* @in es - explainstate information
* @in queryString - prepared sql statement in pbe or simple queryString
* @out - void
*/
static void ExecFQSQueryinDn(ExplainState* es, const char* queryString)
{
ListCell* lc = NULL;
foreach (lc, u_sess->exec_cxt.remotequery_list) {
RemoteQuery* rq = (RemoteQuery*)lfirst(lc);
int dnnum = 0;
StringInfo* rs = SendExplainToDNs(es, rq, &dnnum, queryString);
for (int i = 0; i < dnnum; i++) {
appendStringInfo(es->str, "%s\n", rs[i]->data);
}
}
}
* ExplainOnePlan -
* given a planned query, execute it if needed, and then print
* EXPLAIN output
*
* "into" is NULL unless we are explaining the contents of a CreateTableAsStmt,
* in which case executing the query should result in creating that table.
*
* Since we ignore any DeclareCursorStmt that might be attached to the query,
* if you say EXPLAIN ANALYZE DECLARE CURSOR then we'll actually run the
* query. This is different from pre-8.3 behavior but seems more useful than
* not running the query. No cursor will be created, however.
*
* This is exported because it's called back from prepare.c in the
* EXPLAIN EXECUTE case, and because an index advisor plugin would need
* to call it.
*/
void ExplainOnePlan(
PlannedStmt* plannedstmt,
IntoClause* into,
ExplainState* es,
const char* queryString,
DestReceiver *dest,
ParamListInfo params)
{
QueryDesc* queryDesc = NULL;
instr_time starttime;
instr_time exec_starttime;
double exec_totaltime = 0;
double execrun_totaltime = 0;
double execend_totaltime = 0;
double totaltime = 0;
int eflags;
int instrument_option = 0;
if (es->analyze && es->timing)
instrument_option |= INSTRUMENT_TIMER;
else if (es->analyze)
instrument_option |= INSTRUMENT_ROWS;
if (es->buffers)
instrument_option |= INSTRUMENT_BUFFERS;
#ifndef ENABLE_MULTIPLE_NODES
if (es->cpu)
instrument_option |= INSTRUMENT_CPUS;
#endif
INSTR_TIME_SET_CURRENT(starttime);
* Use a snapshot with an updated command ID to ensure this query sees
* results of any previously executed queries.
*/
PushCopiedSnapshot(GetActiveSnapshot());
UpdateActiveSnapshotCommandId();
queryDesc = CreateQueryDesc(
plannedstmt, queryString, GetActiveSnapshot(), InvalidSnapshot, dest, params, instrument_option);
queryDesc->plannedstmt->instrument_option = instrument_option;
if (es->analyze)
eflags = 0;
else
eflags = EXEC_FLAG_EXPLAIN_ONLY;
if (into != NULL)
eflags |= GetIntoRelEFlags(into);
#ifdef STREAMPLAN
* u_sess->exec_cxt.under_stream_runtime is set to true in ExecInitRemoteQuery() if
* node->is_simple is true. ExecInitRemoteQuery() will be called during
* calling ExecutorStart(queryDesc, eflags).
* light performance is changed to allocate streaminstrumentation firstly,
* and than calling ExecutorStart for ExecInitNode in CN.
*/
#ifdef ENABLE_MULTIPLE_NODES
if (IS_PGXC_COORDINATOR &&
#else
if ((IS_SPQ_COORDINATOR || StreamTopConsumerAmI()) &&
#endif
queryDesc->plannedstmt->is_stream_plan == true &&
check_stream_support() && instrument_option != 0 && u_sess->instr_cxt.global_instr == NULL &&
queryDesc->plannedstmt->num_nodes != 0) {
int dop = queryDesc->plannedstmt->query_dop;
#ifdef USE_SPQ
MemoryContext oldContext = MemoryContextSwitchTo(u_sess->spq_cxt.spq_worker_context);
#endif
u_sess->instr_cxt.global_instr = StreamInstrumentation::InitOnCn(queryDesc, dop);
u_sess->instr_cxt.thread_instr =
u_sess->instr_cxt.global_instr->allocThreadInstrumentation(queryDesc->plannedstmt->planTree->plan_node_id);
{
AutoContextSwitch cxtGuard(SESS_GET_MEM_CXT_GROUP(MEMORY_CONTEXT_OPTIMIZER));
u_sess->instr_cxt.obs_instr = New(CurrentMemoryContext) OBSInstrumentation();
}
#ifdef USE_SPQ
MemoryContextSwitchTo(oldContext);
#endif
}
#endif
INSTR_TIME_SET_CURRENT(exec_starttime);
if (ENABLE_WORKLOAD_CONTROL && (IS_PGXC_COORDINATOR || IS_SINGLE_NODE)) {
u_sess->exec_cxt.need_track_resource =
((unsigned int)eflags & EXEC_FLAG_EXPLAIN_ONLY) ? false : WLMNeedTrackResource(queryDesc);
}
ExecutorStart(queryDesc, eflags);
if (u_sess->attr.attr_common.max_datanode_for_plan > 0 && IS_PGXC_COORDINATOR && !IsConnFromCoord() &&
!es->analyze && es->is_explain_gplan) {
GetNodeList(queryDesc);
}
exec_totaltime += elapsed_time(&exec_starttime);
#ifdef STREAMPLAN
* Init planinfo when:
* 1. explain_perf_mode is not EXPLAIN_NORMAL and it`s a stream plan.
* 2. For EXPLAIN PLAN, we still collect plan info.
*/
if ((t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && IS_STREAM_PLAN &&
u_sess->exec_cxt.under_stream_runtime) ||
es->plan) {
es->planinfo = (PlanInformation*)palloc0(sizeof(PlanInformation));
es->planinfo->init(es, plannedstmt->num_plannodes, plannedstmt->num_nodes, (plannedstmt->query_mem[0] > 0));
} else
t_thrd.explain_cxt.explain_perf_mode = EXPLAIN_NORMAL;
#endif
if (ENABLE_WORKLOAD_CONTROL && (IS_PGXC_COORDINATOR || IS_SINGLE_NODE)) {
if (!((unsigned int)eflags & EXEC_FLAG_EXPLAIN_ONLY)) {
WLMInitQueryPlan(queryDesc);
dywlm_client_manager(queryDesc);
} else if (DY_MEM_ADJ(queryDesc->plannedstmt)) {
bool use_tenant = (u_sess->wlm_cxt->wlm_params.rpdata.rpoid != DEFAULT_POOL_OID);
CalculateQueryMemMain(queryDesc->plannedstmt, use_tenant, true);
}
}
if (es->analyze) {
ScanDirection dir;
if (into != NULL && into->skipData) {
dir = NoMovementScanDirection;
es->sql_execute = false;
} else
dir = ForwardScanDirection;
if (u_sess->exec_cxt.need_track_resource) {
ExplainNodeFinish(queryDesc->planstate, queryDesc->plannedstmt, 0.0, true);
}
INSTR_TIME_SET_CURRENT(exec_starttime);
ExecutorRun(queryDesc, dir, 0L);
execrun_totaltime += elapsed_time(&exec_starttime);
ExecutorFinish(queryDesc);
INSTR_TIME_SET_CURRENT(exec_starttime);
ExecRemoteprocessPlan(queryDesc->estate);
execend_totaltime += elapsed_time(&exec_starttime);
if (u_sess->exec_cxt.need_track_resource &&
u_sess->attr.attr_resource.resource_track_level == RESOURCE_TRACK_OPERATOR) {
List* issueResults = PlanAnalyzerOperator(queryDesc, queryDesc->planstate);
if (issueResults != NIL) {
RecordQueryPlanIssues(issueResults);
}
}
if (u_sess->exec_cxt.need_track_resource) {
u_sess->instr_cxt.can_record_to_table = true;
ExplainNodeFinish(queryDesc->planstate, queryDesc->plannedstmt, GetCurrentTimestamp(), false);
}
totaltime += elapsed_time(&starttime);
}
#ifndef ENABLE_MULTIPLE_NODES
SetNullPrediction(queryDesc->planstate);
if (es->opt_model_name != NULL && PredictorIsValid((const char*)es->opt_model_name)) {
char* file_name = PreModelPredict(queryDesc->planstate, queryDesc->plannedstmt);
ModelPredictForExplain(queryDesc->planstate, file_name, (const char*)es->opt_model_name);
}
#endif
ExplainOpenGroup("Query", NULL, true, es);
if (IS_PGXC_DATANODE && u_sess->attr.attr_sql.enable_opfusion == true &&
es->format == EXPLAIN_FORMAT_TEXT && u_sess->attr.attr_sql.enable_hypo_index == false) {
FusionType type = OpFusion::getFusionType(NULL, params, list_make1(queryDesc->plannedstmt));
if (!es->is_explain_gplan)
type = NOBYPASS_NO_CPLAN;
if (type < BYPASS_OK && type > NONE_FUSION) {
appendStringInfo(es->str, "[Bypass]\n");
} else if (u_sess->attr.attr_sql.opfusion_debug_mode == BYPASS_LOG) {
appendStringInfo(es->str, "[No Bypass]");
const char* bypass_reason = getBypassReason(type);
appendStringInfo(es->str, "reason: %s.\n", bypass_reason);
}
}
ExplainPrintPlan(es, queryDesc);
if (es->post_str != NULL) {
appendStringInfo(es->str, "%s\n", es->post_str->data);
DestroyStringInfo(es->post_str);
es->post_str = NULL;
}
if (es->plan && es->planinfo != NULL) {
es->planinfo->m_planTableData->set_plan_table_ids(queryDesc->plannedstmt->queryId, es);
es->planinfo->m_planTableData->insert_plan_table_tuple();
}
#ifdef STREAMPLAN
if (u_sess->instr_cxt.global_instr && u_sess->instr_cxt.global_instr->needTrack() &&
u_sess->instr_cxt.global_instr->isTrack()) {
if (es->format == EXPLAIN_FORMAT_TEXT && t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL)
show_track_time_info(es);
}
if (u_sess->instr_cxt.obs_instr) {
u_sess->instr_cxt.obs_instr->insertData(queryDesc->plannedstmt->queryId);
}
#endif
if (es->analyze) {
ResultRelInfo* rInfo = NULL;
bool show_relname = false;
int numrels = queryDesc->estate->es_num_result_relations;
List* targrels = queryDesc->estate->es_trig_target_relations;
int nr;
ListCell* l = NULL;
ExplainOpenGroup("Triggers", "Triggers", false, es);
show_relname = (numrels > 1 || targrels != NIL);
rInfo = queryDesc->estate->es_result_relations;
for (nr = 0; nr < numrels; rInfo++, nr++)
report_triggers(rInfo, show_relname, es);
foreach (l, targrels) {
rInfo = (ResultRelInfo*)lfirst(l);
report_triggers(rInfo, show_relname, es);
}
ExplainCloseGroup("Triggers", "Triggers", false, es);
}
if (range_table_walker(
plannedstmt->rtable, (bool (*)())ContainRlsQualInRteWalker, NULL, QTW_EXAMINE_RTES | QTW_IGNORE_DUMMY)) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL
&& es->planinfo->m_detailInfo != NULL) {
appendStringInfo(es->planinfo->m_detailInfo->info_str,
"Notice: This query is influenced by row level security feature\n");
}
ExplainPropertyText("Notice", "This query is influenced by row level security feature", es);
}
if (u_sess->exec_cxt.remotequery_list != NIL) {
if (u_sess->attr.attr_common.max_datanode_for_plan > 0 && IS_PGXC_COORDINATOR && !IsConnFromCoord() &&
!es->analyze) {
ExecFQSQueryinDn(es, queryString);
}
u_sess->exec_cxt.remotequery_list = NIL;
}
if (es->es_frs.node_num > 0) {
appendStringInfo(es->str, "%s\n", es->es_frs.str->data);
pfree_ext(es->es_frs.str->data);
}
* Close down the query and free resources. Include time for this in the
* total runtime (although it should be pretty minimal).
*/
INSTR_TIME_SET_CURRENT(starttime);
ExecutorEnd(queryDesc);
execend_totaltime += elapsed_time(&starttime);
FreeQueryDesc(queryDesc);
PopActiveSnapshot();
if (es->analyze)
CommandCounterIncrement();
totaltime += elapsed_time(&starttime);
if (es->analyze) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL) {
if (MEMORY_TRACKING_QUERY_PEAK)
appendStringInfo(es->str, "Total runtime: %.3f ms, Peak Memory: %ld KB\n", 1000.0 * totaltime,
(int64)(t_thrd.utils_cxt.peakedBytesInQueryLifeCycle/1024));
else
appendStringInfo(es->str, "Total runtime: %.3f ms\n", 1000.0 * totaltime);
}
else if (es->planinfo != NULL && es->planinfo->m_query_summary) {
const char *process_role = "Datanode";
if (IS_PGXC_COORDINATOR) {
process_role = "Coordinator";
}
appendStringInfo(es->planinfo->m_query_summary->info_str, "%s executor start time: %.3f ms\n",
process_role, 1000.0 * exec_totaltime);
appendStringInfo(es->planinfo->m_query_summary->info_str, "%s executor run time: %.3f ms\n",
process_role, 1000.0 * execrun_totaltime);
appendStringInfo(es->planinfo->m_query_summary->info_str, "%s executor end time: %.3f ms\n",
process_role, 1000.0 * execend_totaltime);
if (es->planinfo->m_runtimeinfo && es->planinfo->m_query_summary->m_size) {
long spacePeakKb = (es->planinfo->m_query_summary->m_size + 1023) / 1024;
appendStringInfo(es->planinfo->m_query_summary->info_str, "Total network : %ldkB\n", spacePeakKb);
}
es->planinfo->m_query_summary->m_size = 0;
if (u_sess->analyze_cxt.autoanalyze_timeinfo && u_sess->analyze_cxt.autoanalyze_timeinfo->len > 0) {
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Autoanalyze runtime: %s\n",
u_sess->analyze_cxt.autoanalyze_timeinfo->data);
}
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Planner runtime: %.3f ms\n",
1000.0 * plannedstmt->plannertime);
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Plan size: %d byte\n",
u_sess->instr_cxt.plan_size);
appendStringInfo(es->planinfo->m_query_summary->info_str, "Query Id: %lu\n", u_sess->debug_query_id);
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Total runtime: %.3f ms\n", 1000.0 * totaltime);
}
} else
ExplainPropertyFloat("Total Runtime", 1000.0 * totaltime, 3, es);
}
output_hint_warning(plannedstmt->plan_hint_warning, WARNING);
if (es->verbose)
output_noanalyze_rellist_to_log(WARNING);
ExplainCloseGroup("Query", NULL, true, es);
}
* ExplainPrintPlan -
* convert a QueryDesc's plan tree to text and append it to es->str
*
* The caller should have set up the options fields of *es, as well as
* initializing the output buffer es->str. Other fields in *es are
* initialized here.
*
* NB: will not work on utility statements
*/
void ExplainPrintPlan(ExplainState* es, QueryDesc* queryDesc)
{
AssertEreport(queryDesc->plannedstmt != NULL, MOD_EXECUTOR, "unexpect null value");
es->pstmt = queryDesc->plannedstmt;
es->rtable = queryDesc->plannedstmt->rtable;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL) {
es->planinfo->m_query_id = queryDesc->plannedstmt->queryId;
if (es->analyze && es->planinfo->m_staticInfo)
show_peak_memory(es, es->planinfo->m_planInfo->m_size);
ExplainNode<true>(queryDesc->planstate, NIL, NULL, NULL, es);
if (u_sess->attr.attr_sql.plan_mode_seed != OPTIMIZE_PLAN) {
appendStringInfo(es->planinfo->m_query_summary->info_str, "Random seed: %u\n", get_inital_plan_seed());
}
if (es->pstmt->query_mem[0] != 0) {
if (es->pstmt->assigned_query_mem[0] > 0)
appendStringInfo(es->planinfo->m_query_summary->info_str,
"System available mem: %dKB\n",
es->pstmt->assigned_query_mem[0]);
if (es->pstmt->assigned_query_mem[1] > 0)
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Query Max mem: %dKB\n", es->pstmt->assigned_query_mem[1]);
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Query estimated mem: %dKB\n", es->pstmt->query_mem[0]);
}
if (es->pstmt->is_dynmaic_smp) {
if (g_instance.wlm_cxt->dynamic_workload_inited) {
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Avail/Max core: %d/%d\n",
es->pstmt->dynsmp_avail_cpu,
es->pstmt->dynsmp_max_cpu);
appendStringInfo(es->planinfo->m_query_summary->info_str, "Cpu util: %d\n", es->pstmt->dynsmp_cpu_util);
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Active statement: %d\n",
es->pstmt->dynsmp_active_statement);
} else {
appendStringInfo(es->planinfo->m_query_summary->info_str, "Avail/Max core: NA/NA\n");
appendStringInfo(es->planinfo->m_query_summary->info_str, "Cpu util: NA\n");
appendStringInfo(es->planinfo->m_query_summary->info_str, "Active statement: NA\n");
}
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Query estimated cpu: %.1lf\n",
es->pstmt->dynsmp_query_estimate_cpu_usge);
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Mem allowed dop: %d\n", es->pstmt->dynsmp_dop_mem_limit);
if (es->pstmt->dynsmp_min_non_spill_dop == DYNMSP_ALREADY_SPILLED) {
appendStringInfo(es->planinfo->m_query_summary->info_str, "Min non-spill dop: already spilled\n");
} else if (es->pstmt->dynsmp_min_non_spill_dop == DYNMSP_SPILL_IF_USE_HIGHER_DOP) {
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Min non-spill dop: spill if use higher dop\n");
} else {
appendStringInfo(es->planinfo->m_query_summary->info_str,
"Min non-spill dop: %d\n",
es->pstmt->dynsmp_min_non_spill_dop);
}
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Initial dop: %d\n", es->pstmt->dynsmp_plan_original_dop);
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Final dop: %d\n", es->pstmt->dynsmp_plan_optimal_dop);
}
} else
ExplainNode<false>(queryDesc->planstate, NIL, NULL, NULL, es);
}
* ExplainQueryText -
* add a "Query Text" node that contains the actual text of the query
*
* The caller should have set up the options fields of *es, as well as
* initializing the output buffer es->str.
*
*/
void ExplainQueryText(ExplainState* es, QueryDesc* queryDesc)
{
if (queryDesc->sourceText)
ExplainPropertyText("Query Text", queryDesc->sourceText, es);
}
* Explain a list of children of a ExtensiblePlan.
*/
static void ExplainExtensibleChildren(ExtensiblePlanState* planState, List* ancestors, ExplainState* es)
{
ListCell* cell = NULL;
const char* label = (list_length(planState->extensible_ps) != 1 ? "children" : "child");
foreach (cell, planState->extensible_ps)
ExplainNode<false>((PlanState*)lfirst(cell), ancestors, label, NULL, es);
}
* ExplainOneQueryForStatistic -
* print out the execution plan for one Query for wlm statistics
* in - queryDesc: query plan description.
*/
void ExplainOneQueryForStatistics(QueryDesc* queryDesc)
{
ExplainState es;
ExplainInitState(&es);
es.costs = true;
es.nodes = false;
es.format = EXPLAIN_FORMAT_TEXT;
int before_explain_perf_mode = t_thrd.explain_cxt.explain_perf_mode;
t_thrd.explain_cxt.explain_perf_mode = EXPLAIN_NORMAL;
if (queryDesc->plannedstmt == NULL || queryDesc->planstate == NULL)
return;
es.pstmt = queryDesc->plannedstmt;
es.rtable = queryDesc->plannedstmt->rtable;
int max_plan_id = queryDesc->plannedstmt->num_plannodes;
int wlm_statistics_plan_max_digit = get_digit(max_plan_id);
es.wlm_statistics_plan_max_digit = &wlm_statistics_plan_max_digit;
#ifndef ENABLE_MULTIPLE_NODES
SetNullPrediction(queryDesc->planstate);
#endif
StreamInstrumentation* oldInstr = u_sess->instr_cxt.global_instr;
u_sess->instr_cxt.global_instr = NULL;
u_sess->exec_cxt.under_auto_explain = true;
appendStringInfo(es.str, "Coordinator Name: %s\n", g_instance.attr.attr_common.PGXCNodeName);
ExplainNode<false>(queryDesc->planstate, NIL, NULL, NULL, &es);
u_sess->exec_cxt.under_auto_explain = false;
u_sess->instr_cxt.global_instr = oldInstr;
AutoContextSwitch memSwitch(g_instance.wlm_cxt->query_resource_track_mcxt);
t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->query_plan = (char*)palloc0(es.str->len + 1 + 1);
errno_t rc =
memcpy_s(t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->query_plan, es.str->len, es.str->data, es.str->len);
securec_check(rc, "\0", "\0");
t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->query_plan[es.str->len] = '\n';
t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->query_plan[es.str->len + 1] = '\0';
t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->plan_size = es.str->len + 1 + 1;
t_thrd.explain_cxt.explain_perf_mode = before_explain_perf_mode;
}
* report_triggers -
* report execution stats for a single relation's triggers
*/
static void report_triggers(ResultRelInfo* rInfo, bool show_relname, ExplainState* es)
{
int nt;
if (!rInfo->ri_TrigDesc || !rInfo->ri_TrigInstrument)
return;
for (nt = 0; nt < rInfo->ri_TrigDesc->numtriggers; nt++) {
Trigger* trig = rInfo->ri_TrigDesc->triggers + nt;
Instrumentation* instr = rInfo->ri_TrigInstrument + nt;
char* relname = NULL;
char* conname = NULL;
InstrEndLoop(instr);
* We ignore triggers that were never invoked; they likely aren't
* relevant to the current query type.
*/
if (instr->ntuples == 0)
continue;
ExplainOpenGroup("Trigger", NULL, true, es);
relname = RelationGetRelationName(rInfo->ri_RelationDesc);
if (OidIsValid(trig->tgconstraint))
conname = get_constraint_name(trig->tgconstraint);
* In text format, we avoid printing both the trigger name and the
* constraint name unless VERBOSE is specified. In non-text formats
* we just print everything.
*/
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (es->verbose || conname == NULL)
appendStringInfo(es->str, "Trigger %s", trig->tgname);
else
appendStringInfoString(es->str, "Trigger");
if (conname != NULL)
appendStringInfo(es->str, " for constraint %s", conname);
if (show_relname)
appendStringInfo(es->str, " on %s", relname);
appendStringInfo(es->str, ": time=%.3f calls=%.0f\n", 1000.0 * instr->total, instr->ntuples);
} else {
ExplainPropertyText("Trigger Name", trig->tgname, es);
if (conname != NULL)
ExplainPropertyText("Constraint Name", conname, es);
ExplainPropertyText("Relation", relname, es);
ExplainPropertyFloat("Time", 1000.0 * instr->total, 3, es);
ExplainPropertyFloat("Calls", instr->ntuples, 0, es);
}
if (conname != NULL)
pfree_ext(conname);
ExplainCloseGroup("Trigger", NULL, true, es);
}
}
double elapsed_time(instr_time* starttime)
{
instr_time endtime;
INSTR_TIME_SET_CURRENT(endtime);
INSTR_TIME_SUBTRACT(endtime, *starttime);
return INSTR_TIME_GET_DOUBLE(endtime);
}
static void show_bucket_info(PlanState* planstate, ExplainState* es, bool is_pretty)
{
Scan* scanplan = (Scan*)planstate->plan;
BucketInfo* bucket_info = scanplan->bucketInfo;
int selected_buckets = list_length(bucket_info->buckets);
if (selected_buckets == 0) {
selected_buckets = BUCKETDATALEN;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty) {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str,
"Selected Buckets %d of %d : %s\n",
selected_buckets,
BUCKETDATALEN,
bucketInfoToString(bucket_info));
} else {
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
}
appendStringInfo(es->str,
"Selected Buckets %d of %d : %s\n",
selected_buckets,
BUCKETDATALEN,
bucketInfoToString(bucket_info));
}
}
}
typedef struct PrintFlags {
bool print_number;
bool print_twodots;
bool print_comma;
bool print_prev;
} PrintFlags;
static void set_print_flags(int i, bool showPrev, bool isLast, PrintFlags* flags)
{
if (i == 0) {
flags->print_prev = false;
flags->print_number = true;
flags->print_twodots = false;
flags->print_comma = false;
} else if (showPrev) {
if (isLast) {
flags->print_prev = false;
flags->print_number = true;
flags->print_twodots = true;
flags->print_comma = false;
} else {
flags->print_prev = false;
flags->print_number = false;
flags->print_twodots = false;
flags->print_comma = false;
}
} else {
if (!flags->print_number)
flags->print_prev = true;
else
flags->print_prev = false;
flags->print_number = true;
flags->print_twodots = false;
flags->print_comma = true;
}
}
static void add_pruning_info_nums(StringInfo dest, PrintFlags flags, int partID, int partID_prev)
{
if (flags.print_prev) {
appendStringInfoString(dest, "..");
appendStringInfo(dest, "%d", partID_prev + 1);
}
if (flags.print_twodots) {
appendStringInfoString(dest, "..");
} else if (flags.print_comma) {
appendStringInfoString(dest, ",");
}
if (flags.print_number) {
appendStringInfo(dest, "%d", partID + 1);
}
}
* Show number of subpartitions selected for each subpartition.
* If subpartitions is not pruned at all, show "ALL".
* The caller should call DestroyStringInfo to free allocated memory.
*/
static StringInfo get_subpartition_pruning_info(Scan* scanplan, List* rtable)
{
PruningResult* pr = scanplan->pruningInfo;
StringInfo strif = makeStringInfo();
ListCell* lc = NULL;
bool all = true;
RangeTblEntry* rte = rt_fetch(scanplan->scanrelid, rtable);
Relation rel = heap_open(rte->relid, NoLock);
List* subpartList = RelationGetSubPartitionOidListList(rel);
int idx = 0;
foreach (lc, pr->ls_selectedSubPartitions) {
idx++;
if (idx > list_length(subpartList)) {
break;
}
SubPartitionPruningResult* spr = (SubPartitionPruningResult*)lfirst(lc);
int selected = list_length(spr->ls_selectedSubPartitions);
if (spr->partSeq >= list_length(subpartList)) {
all = false;
} else {
int count = list_length((List*)list_nth(subpartList, spr->partSeq));
all &= (selected == count);
}
appendStringInfo(strif, "%d:%d", spr->partSeq + 1, selected);
if (lc != list_tail(pr->ls_selectedSubPartitions)) {
appendStringInfo(strif, ", ");
}
}
if (all) {
resetStringInfo(strif);
appendStringInfo(strif, "ALL");
}
ReleaseSubPartitionOidList(&subpartList);
heap_close(rel, NoLock);
return strif;
}
* Show subpartition pruning information in the form of:
* Selected Partitions: 1,3
* Selected Subpartitions: 1:1,3:1
* Which means select first subpartition of 1 and 3 partition
* Note that pretty mode is not supported in this version, since subpartitions can only
* be created in Cent deployment for now and pretty is only for stream plans.
*/
static void add_subpartition_pruning_info_text(PlanState* planstate, ExplainState* es)
{
Scan* scanplan = (Scan*)planstate->plan;
PruningResult* pr = scanplan->pruningInfo;
if (pr->ls_selectedSubPartitions == NIL) {
return;
}
StringInfo dest = es->str;
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(dest, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(dest, " | ");
appendStringInfoSpaces(dest, es->indent);
} else {
appendStringInfoSpaces(dest, es->indent * 2);
}
appendStringInfo(dest, "Selected Subpartitions: ");
if (scanplan->itrs <= 0) {
appendStringInfo(dest, "NONE");
} else if (scanplan->pruningInfo->expr != NULL) {
appendStringInfo(dest, "PART");
} else {
StringInfo strif = get_subpartition_pruning_info(scanplan, es->rtable);
appendStringInfoString(dest, strif->data);
DestroyStringInfo(strif);
}
appendStringInfoChar(dest, '\n');
}
static void add_subpartition_pruning_info_others(PlanState* planstate, ExplainState* es)
{
Scan* scanplan = (Scan*)planstate->plan;
PruningResult* pr = scanplan->pruningInfo;
if (pr->ls_selectedSubPartitions == NIL) {
return;
}
if (scanplan->itrs <= 0) {
ExplainPropertyText("Selected Subpartitions", "NONE", es);
} else if (scanplan->pruningInfo->expr != NULL) {
ExplainPropertyText("Selected Subpartitions", "PART", es);
} else {
StringInfo strif = get_subpartition_pruning_info(scanplan, es->rtable);
ExplainPropertyText("Selected Subpartitions", strif->data, es);
DestroyStringInfo(strif);
}
}
static void show_pruning_info(PlanState* planstate, ExplainState* es, bool is_pretty)
{
Scan* scanplan = (Scan*)planstate->plan;
PrintFlags flags = {false, false, false, false};
int partID = 0;
int partID_prev = 0;
if (es->format == EXPLAIN_FORMAT_TEXT) {
int i = 0;
if (is_pretty == false) {
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
}
appendStringInfo(es->str, "Selected Partitions: ");
} else {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str, "Selected Partitions: ");
}
if (scanplan->itrs <= 0) {
if (is_pretty == false)
appendStringInfo(es->str, "NONE");
else
appendStringInfo(es->planinfo->m_detailInfo->info_str, "NONE");
} else if (scanplan->pruningInfo->expr != NULL) {
appendStringInfo(es->str, "%s", "PART");
} else {
ListCell* cell = NULL;
List* part_seqs = scanplan->pruningInfo->ls_rangeSelectedPartitions;
AssertEreport(scanplan->itrs == scanplan->pruningInfo->ls_rangeSelectedPartitions->length,
MOD_EXECUTOR,
"unexpect list length");
foreach (cell, part_seqs) {
if (i > 0) {
partID_prev = partID;
}
partID = lfirst_int(cell);
if (partID < 0)
break;
set_print_flags(i, partID - partID_prev == 1, i == scanplan->itrs - 1, &flags);
if (is_pretty) {
add_pruning_info_nums(es->planinfo->m_detailInfo->info_str, flags, partID, partID_prev);
} else {
add_pruning_info_nums(es->str, flags, partID, partID_prev);
}
i++;
}
}
if (is_pretty)
appendStringInfoChar(es->planinfo->m_detailInfo->info_str, '\n');
else
appendStringInfoChar(es->str, '\n');
if (scanplan->itrs > 0 && scanplan->pruningInfo->expr == NULL) {
add_subpartition_pruning_info_text(planstate, es);
}
} else {
if (scanplan->itrs <= 0) {
ExplainPropertyText("Selected Partitions", "NONE", es);
} else if (scanplan->pruningInfo->expr != NULL) {
ExplainPropertyText("Selected Partitions", "PART", es);
} else {
int i = 0;
StringInfo strif;
ListCell* cell = NULL;
List* part_seqs = scanplan->pruningInfo->ls_rangeSelectedPartitions;
AssertEreport(scanplan->itrs == scanplan->pruningInfo->ls_rangeSelectedPartitions->length,
MOD_EXECUTOR,
"unexpect list length");
strif = makeStringInfo();
foreach (cell, part_seqs) {
if (i > 0) {
partID_prev = partID;
}
partID = lfirst_int(cell);
if (partID < 0)
break;
set_print_flags(i, partID - partID_prev == 1, i == scanplan->itrs - 1, &flags);
add_pruning_info_nums(strif, flags, partID, partID_prev);
i++;
}
ExplainPropertyText("Selected Partitions", strif->data, es);
pfree_ext(strif->data);
pfree_ext(strif);
add_subpartition_pruning_info_others(planstate, es);
}
}
}
static void ExplainNodePartition(const Plan* plan, ExplainState* es)
{
int flag = 0;
switch (nodeTag(plan->lefttree)) {
case T_SeqScan:
if (((SeqScan*)plan->lefttree)->pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
#ifdef USE_SPQ
case T_SpqSeqScan:
if (((SpqSeqScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_SpqIndexScan:
if (((SpqIndexScan*)plan->lefttree)->scan.scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_SpqIndexOnlyScan:
if (((SpqIndexOnlyScan*)plan->lefttree)->scan.scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_SpqBitmapHeapScan:
if (((SpqBitmapHeapScan*)plan->lefttree)->scan.scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
#endif
case T_IndexScan:
if (((IndexScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_IndexOnlyScan:
if (((IndexOnlyScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_BitmapIndexScan:
if (((BitmapIndexScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_BitmapHeapScan:
if (((BitmapHeapScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_CStoreScan:
if (((CStoreScan*)plan->lefttree)->pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
case T_AnnIndexScan:
if (((AnnIndexScan*)plan->lefttree)->scan.pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
if (((IMCStoreScan*)plan->lefttree)->pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
if (((TsStoreScan*)plan->lefttree)->pruningInfo->expr != NULL) {
appendStringInfo(es->str, "Iterations: %s", "PART");
flag = 1;
}
break;
#endif
default:
appendStringInfo(es->str, "Iterations: %d", ((PartIterator*)plan)->itrs);
flag = 1;
break;
}
if (flag == 0) {
appendStringInfo(es->str, "Iterations: %d", ((PartIterator*)plan)->itrs);
}
}
static bool GetSubPartitionIterations(const Plan* plan, const ExplainState* es, int* cnt)
{
*cnt = 0;
const Plan* curPlan = plan;
switch (nodeTag(curPlan->lefttree)) {
case T_RowToVec: {
RowToVec* rowToVecPlan = (RowToVec*)curPlan->lefttree;
Plan* scanPlan = (Plan*)rowToVecPlan->plan.lefttree;
if (!(IsA(scanPlan, Scan) || IsA(scanPlan, SeqScan) || IsA(scanPlan, IndexOnlyScan) ||
IsA(scanPlan, IndexScan) || IsA(scanPlan, BitmapHeapScan) || IsA(scanPlan, TidScan) ||
IsA(scanPlan, AnnIndexScan))) {
break;
}
curPlan = &rowToVecPlan->plan;
}
case T_SeqScan:
#ifdef USE_SPQ
case T_SpqSeqScan:
case T_SpqIndexScan:
case T_SpqIndexOnlyScan:
case T_SpqBitmapHeapScan:
#endif
case T_IndexScan:
case T_IndexOnlyScan:
case T_AnnIndexScan:
case T_BitmapIndexScan:
case T_BitmapHeapScan:
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
case T_TidScan: {
PruningResult* pr = ((Scan*)curPlan->lefttree)->pruningInfo;
if (pr == NULL || pr->ls_selectedSubPartitions == NIL || pr->expr != NULL) {
return false;
}
ListCell* lc = NULL;
foreach (lc, pr->ls_selectedSubPartitions) {
SubPartitionPruningResult* spr = (SubPartitionPruningResult*)lfirst(lc);
*cnt += list_length(spr->ls_selectedSubPartitions);
}
return true;
}
default:
return false;
}
return false;
}
#ifndef ENABLE_MULTIPLE_NODES
static void PredAppendInfo(Plan* plan, StringInfoData buf, ExplainState* es)
{
if (es->planinfo->m_planInfo->m_pred_time && plan->pred_total_time >= 0) {
initStringInfo(&buf);
appendStringInfo(&buf, "%.0f", plan->pred_total_time);
es->planinfo->m_planInfo->put(PREDICT_TIME, PointerGetDatum(cstring_to_text(buf.data)));
pfree_ext(buf.data);
}
if (es->planinfo->m_planInfo->m_pred_row && plan->pred_rows >= 0) {
es->planinfo->m_planInfo->put(PREDICT_ROWS,
DirectFunctionCall1(dround,
Float8GetDatum(plan->pred_rows)));
}
if (es->planinfo->m_planInfo->m_pred_mem && plan->pred_max_memory >= 0) {
es->planinfo->m_planInfo->put(PREDICT_MEMORY,
DirectFunctionCall1(pg_size_pretty,
Int64GetDatum(plan->pred_max_memory)));
}
}
static void PredGetInfo(Plan* plan, ExplainState* es)
{
if (plan->pred_startup_time >= 0) {
ExplainPropertyFloat("Pred Startup Time", plan->pred_startup_time, 2, es);
}
if (plan->pred_total_time >= 0) {
ExplainPropertyFloat("Pred Total Time", plan->pred_total_time, 2, es);
}
if (plan->pred_rows >= 0) {
ExplainPropertyFloat("Pred Rows", plan->pred_rows, 0, es);
}
if (plan->pred_max_memory >= 0) {
ExplainPropertyLong("Pred Peak Memory", plan->pred_max_memory, es);
}
}
#endif
* ExplainNode -
* Appends a description of a plan tree to es->str
*
* planstate points to the executor state node for the current plan node.
* We need to work from a PlanState node, not just a Plan node, in order to
* get at the instrumentation data (if any) as well as the list of subplans.
*
* ancestors is a list of parent PlanState nodes, most-closely-nested first.
* These are needed in order to interpret PARAM_EXEC Params.
*
* relationship describes the relationship of this plan node to its parent
* (eg, "Outer", "Inner"); it can be null at top level. plan_name is an
* optional name to be attached to the node.
*
* In text format, es->indent is controlled in this function since we only
* want it to change at plan-node boundaries. In non-text formats, es->indent
* corresponds to the nesting depth of logical output groups, and therefore
* is controlled by ExplainOpenGroup/ExplainCloseGroup.
*/
template <bool is_pretty>
static void ExplainNode(
PlanState* planstate, List* ancestors, const char* relationship, const char* plan_name, ExplainState* es)
{
Plan* plan = planstate->plan;
char* pname = NULL;
char* sname = NULL;
char* strategy = NULL;
char* operation = NULL;
int save_indent = es->indent;
int save_pindent = es->pindent;
bool haschildren = false;
int plan_node_id = planstate->plan->plan_node_id;
int parentid = planstate->plan->parent_node_id;
StringInfo tmpName = nullptr;
bool from_datanode = false;
bool old_dn_flag = false;
#ifdef USE_SPQ
int current_id_bak = es->current_id;
#endif
char* pt_operation = NULL;
char* pt_options = NULL;
const char* pt_index_name = NULL;
const char* pt_index_owner = NULL;
if (is_pretty) {
if (plan->plan_node_id == 0)
goto runnext;
tmpName = &es->planinfo->m_planInfo->m_pname;
resetStringInfo(tmpName);
}
if (is_pretty) {
es->planinfo->set_id(plan_node_id);
es->planinfo->m_planInfo->put(PLANID, plan_node_id);
if (es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->m_plan_node_id = plan_node_id;
es->planinfo->m_runtimeinfo->put(-1, -1, PLAN_PARENT_ID, parentid);
es->planinfo->m_runtimeinfo->put(-1, -1, QUERY_ID, es->planinfo->m_query_id);
es->planinfo->m_runtimeinfo->put(-1, -1, PLAN_NODE_ID, plan_node_id);
es->planinfo->m_runtimeinfo->put(-1, -1, PLAN_TYPE, nodeTag(plan));
from_datanode = es->from_dn;
es->planinfo->m_runtimeinfo->put(-1, -1, FROM_DATANODE, from_datanode);
}
if (es->analyze && (IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery))) {
show_dn_executor_time(es, plan_node_id, DN_START_TIME);
show_dn_executor_time(es, plan_node_id, DN_RUN_TIME);
show_dn_executor_time(es, plan_node_id, DN_END_TIME);
}
}
GetPlanNodePlainText(plan, &pname, &sname, &strategy, &operation, &pt_operation, &pt_options);
ExplainOpenGroup("Plan", relationship ? NULL : "Plan", true, es);
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty) {
appendStringInfoString(tmpName, pname);
} else {
if (es->wlm_statistics_plan_max_digit) {
show_wlm_explain_name(es, plan_name, pname, plan_node_id);
} else {
if (plan_name != NULL) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "%s\n", plan_name);
es->indent++;
}
if (es->indent) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, "-> ");
es->indent += 2;
}
appendStringInfoString(es->str, pname);
es->indent++;
}
}
} else {
ExplainPropertyText("Node Type", sname, es);
if (strategy != NULL)
ExplainPropertyText("Strategy", strategy, es);
if (operation != NULL)
ExplainPropertyText("Operation", operation, es);
if (relationship != NULL)
ExplainPropertyText("Parent Relationship", relationship, es);
if (plan_name != NULL)
ExplainPropertyText("Subplan Name", plan_name, es);
}
switch (nodeTag(plan)) {
case T_SeqScan:
#ifdef USE_SPQ
case T_SpqSeqScan:
case T_SpqBitmapHeapScan:
#endif
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
#endif
case T_BitmapHeapScan:
case T_CStoreIndexHeapScan:
case T_TidScan:
case T_TidRangeScan:
case T_SubqueryScan:
case T_VecSubqueryScan:
case T_FunctionScan:
case T_ValuesScan:
case T_CteScan:
case T_WorkTableScan:
case T_ForeignScan:
case T_VecForeignScan:
if (((Scan *) plan)->scanrelid > 0) {
ExplainScanTarget((Scan*)plan, es);
}
break;
case T_TrainModel:
appendStringInfo(es->str, " - %s", sname);
break;
case T_ExtensiblePlan:
if (((Scan*)plan)->scanrelid > 0)
ExplainScanTarget((Scan*)plan, es);
break;
#ifdef USE_SPQ
case T_Stream: {
es->current_id = ((Stream*)plan)->streamID;
}
break;
case T_ShareInputScan: {
ShareInputScan *sisc = (ShareInputScan *) plan;
int slice_id = es->current_id;
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, " (%s; slice%d; share%d; producer:%d)",
(sisc->is_producer ? "Producer" : "Consumer"),
slice_id, sisc->share_id, sisc->producer_slice_id);
} else {
ExplainPropertyText("Identity", (sisc->is_producer ? "Producer" : "Consumer"), es);
ExplainPropertyInteger("Producer ID", sisc->producer_slice_id, es);
ExplainPropertyInteger("Share ID", sisc->share_id, es);
ExplainPropertyInteger("Slice ID", slice_id, es);
}
}
break;
#endif
#ifdef PGXC
case T_RemoteQuery:
case T_VecRemoteQuery:
#ifdef USE_SPQ
es->current_id = ((RemoteQuery*)plan)->streamID;
#endif
ExplainExecNodes(((RemoteQuery*)plan)->exec_nodes, es);
#ifdef STREAMPLAN
if (!IS_STREAM)
#endif
ExplainScanTarget((Scan*)plan, es);
break;
#endif
#ifdef USE_SPQ
case T_SpqIndexScan:
#endif
case T_IndexScan: {
IndexScan* indexscan = (IndexScan*)plan;
ExplainIndexScanDetails(indexscan->indexid, indexscan->indexorderdir, es);
ExplainScanTarget((Scan*)indexscan, es);
pt_index_name = explain_get_index_name(indexscan->indexid);
pt_index_owner = get_namespace_name(get_rel_namespace(indexscan->indexid));
} break;
#ifdef USE_SPQ
case T_SpqIndexOnlyScan:
#endif
case T_IndexOnlyScan: {
IndexOnlyScan* indexonlyscan = (IndexOnlyScan*)plan;
ExplainIndexScanDetails(indexonlyscan->indexid, indexonlyscan->indexorderdir, es);
ExplainScanTarget((Scan*)indexonlyscan, es);
pt_index_name = explain_get_index_name(indexonlyscan->indexid);
pt_index_owner = get_namespace_name(get_rel_namespace(indexonlyscan->indexid));
} break;
case T_AnnIndexScan: {
AnnIndexScan* annindexscan = (AnnIndexScan*)plan;
ExplainIndexScanDetails(annindexscan->indexid, annindexscan->indexorderdir, es);
ExplainScanTarget((Scan*)annindexscan, es);
pt_index_name = explain_get_index_name(annindexscan->indexid);
pt_index_owner = get_namespace_name(get_rel_namespace(annindexscan->indexid));
} break;
case T_BitmapIndexScan: {
BitmapIndexScan* bitmapindexscan = (BitmapIndexScan*)plan;
const char* indexname = explain_get_index_name(bitmapindexscan->indexid);
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, " on %s", indexname);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
tmpName = &es->planinfo->m_planInfo->m_pname;
appendStringInfo(tmpName, " using %s", indexname);
}
} else {
ExplainPropertyText("Index Name", indexname, es);
}
pt_index_name = indexname;
pt_index_owner = get_namespace_name(get_rel_namespace(bitmapindexscan->indexid));
} break;
case T_CStoreIndexScan: {
CStoreIndexScan* indexscan = (CStoreIndexScan*)plan;
ExplainIndexScanDetails(indexscan->indexid, indexscan->indexorderdir, es);
ExplainScanTarget((Scan*)indexscan, es);
pt_index_name = explain_get_index_name(indexscan->indexid);
pt_index_owner = get_namespace_name(get_rel_namespace(indexscan->indexid));
} break;
case T_CStoreIndexCtidScan: {
CStoreIndexCtidScan* bitmapindexscan = (CStoreIndexCtidScan*)plan;
const char* indexname = explain_get_index_name(bitmapindexscan->indexid);
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfo(es->str, " on %s", indexname);
else
ExplainPropertyText("Index Name", indexname, es);
pt_index_name = indexname;
pt_index_owner = get_namespace_name(get_rel_namespace(bitmapindexscan->indexid));
} break;
case T_ModifyTable:
case T_VecModifyTable:
ExplainModifyTarget((ModifyTable*)plan, es);
break;
case T_NestLoop:
case T_VecNestLoop:
case T_VecMergeJoin:
case T_MergeJoin:
case T_AsofJoin:
case T_VecAsofJoin:
case T_HashJoin:
case T_VecHashJoin: {
const char* jointype = NULL;
switch (((Join*)plan)->jointype) {
case JOIN_INNER:
jointype = pt_options = "Inner";
break;
case JOIN_LEFT:
jointype = pt_options = "Left";
break;
case JOIN_FULL:
jointype = pt_options = "Full";
break;
case JOIN_RIGHT:
jointype = pt_options = "Right";
break;
case JOIN_SEMI:
jointype = pt_options = "Semi";
break;
case JOIN_ANTI:
jointype = pt_options = "Anti";
break;
case JOIN_RIGHT_SEMI:
jointype = pt_options = "Right Semi";
break;
case JOIN_RIGHT_ANTI:
jointype = pt_options = "Right Anti";
break;
case JOIN_LEFT_ANTI_FULL:
jointype = pt_options = "Left Anti Full";
break;
case JOIN_RIGHT_ANTI_FULL:
jointype = pt_options = "Right Anti Full";
break;
#ifdef USE_SPQ
case JOIN_LASJ_NOTIN:
jointype = "Left Anti Semi (Not-In)";
break;
#endif
default:
jointype = pt_options = "?\?\?";
break;
}
if (es->plan && es->planinfo != NULL)
es->planinfo->m_planTableData->set_plan_table_join_options(plan, &pt_options);
if (es->format == EXPLAIN_FORMAT_TEXT) {
* For historical reasons, the join type is interpolated
* into the node type name...
*/
if (is_pretty == false) {
if (((Join*)plan)->jointype != JOIN_INNER) {
appendStringInfo(es->str, " %s Join", jointype);
} else if (!(IsA(plan, NestLoop) || IsA(plan, VecNestLoop))) {
appendStringInfo(es->str, " Join");
}
} else {
if (((Join*)plan)->jointype != JOIN_INNER) {
appendStringInfo(tmpName,
" %s Join (%d, %d)",
jointype,
planstate->lefttree->plan->plan_node_id,
planstate->righttree->plan->plan_node_id);
} else if (!(IsA(plan, NestLoop) || IsA(plan, VecNestLoop))) {
appendStringInfo(tmpName,
" Join (%d,%d)",
planstate->lefttree->plan->plan_node_id,
planstate->righttree->plan->plan_node_id);
} else {
appendStringInfo(tmpName,
" (%d,%d)",
planstate->lefttree->plan->plan_node_id,
planstate->righttree->plan->plan_node_id);
}
}
} else
ExplainPropertyText("Join Type", jointype, es);
} break;
case T_SetOp:
case T_VecSetOp: {
const char* setopcmd = NULL;
switch (((SetOp*)plan)->cmd) {
case SETOPCMD_INTERSECT:
setopcmd = "Intersect";
break;
case SETOPCMD_INTERSECT_ALL:
setopcmd = "Intersect All";
break;
case SETOPCMD_EXCEPT:
setopcmd = "Except";
break;
case SETOPCMD_EXCEPT_ALL:
setopcmd = "Except All";
break;
default:
setopcmd = "?\?\?";
break;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty == false)
appendStringInfo(es->str, " %s", setopcmd);
else
appendStringInfo(tmpName, " %s", setopcmd);
} else
ExplainPropertyText("Command", setopcmd, es);
} break;
case T_PartIterator: {
const char* scandir = NULL;
if (ScanDirectionIsBackward(((PartIterator*)plan)->direction))
scandir = "Scan Backward";
if (scandir != NULL) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty == false) {
appendStringInfo(es->str, " %s", scandir);
} else {
appendStringInfo(tmpName, " %s", scandir);
}
} else {
ExplainPropertyText("Scan Direction", scandir, es);
}
}
} break;
case T_Append:
case T_VecAppend: {
if (is_pretty) {
int nplans = list_length(((Append*)plan)->appendplans);
PlanState** append_planstate = ((AppendState*)planstate)->appendplans;
bool first = true;
appendStringInfo(tmpName, "(");
for (int j = 0; j < nplans; j++) {
if (first == false)
appendStringInfoString(tmpName, ", ");
appendStringInfo(tmpName, "%d", append_planstate[j]->plan->plan_node_id);
first = false;
}
appendStringInfo(tmpName, ")");
}
} break;
case T_RecursiveUnion: {
if (es->format == EXPLAIN_FORMAT_TEXT && is_pretty)
appendStringInfo(tmpName,
" (%d,%d)",
planstate->lefttree->plan->plan_node_id,
planstate->righttree->plan->plan_node_id);
} break;
default:
break;
}
if (GET_CONTROL_PLAN_NODEID(plan) != 0 && u_sess->attr.attr_sql.enable_stream_recursive) {
appendStringInfo(
es->str, " <<ruid:[%d] ctlid:[%d]", GET_RECURSIVE_UNION_PLAN_NODEID(plan), GET_CONTROL_PLAN_NODEID(plan));
if (is_pretty)
appendStringInfo(tmpName,
" <<ruid:[%d] ctlid:[%d]",
GET_RECURSIVE_UNION_PLAN_NODEID(plan),
GET_CONTROL_PLAN_NODEID(plan));
if (plan->is_sync_plannode) {
appendStringInfo(es->str, " (SYNC)");
if (is_pretty)
appendStringInfo(tmpName, " (SYNC)");
}
appendStringInfo(es->str, ">>");
if (is_pretty)
appendStringInfo(tmpName, ">>");
}
if (u_sess->attr.attr_sql.enable_stream_recursive && IsA(plan, Stream) && ((Stream*)plan)->stream_level > 0 &&
plan->recursive_union_plan_nodeid != 0) {
Stream* stream_plan = (Stream*)plan;
appendStringInfo(es->str, " stream_level:%d ", stream_plan->stream_level);
if (is_pretty)
appendStringInfo(tmpName, " stream_level:%d ", stream_plan->stream_level);
}
if (ndp_pushdown_hook) {
show_ndpplugin_statistic(es, planstate, tmpName, is_pretty);
}
if (is_pretty) {
StringInfoData pretty_plan_name;
initStringInfo(&pretty_plan_name);
appendStringInfoSpaces(&pretty_plan_name, es->pindent);
appendStringInfoString(&pretty_plan_name, "-> ");
es->pindent += 2;
appendStringInfoString(&pretty_plan_name, es->planinfo->m_planInfo->m_pname.data);
if (plan_name != NULL) {
appendStringInfoSpaces(&pretty_plan_name, 2);
appendStringInfo(&pretty_plan_name, "[%d, %s]", plan->parent_node_id, plan_name);
}
es->planinfo->m_planInfo->put(PLAN, PointerGetDatum(cstring_to_text(pretty_plan_name.data)));
pfree_ext(pretty_plan_name.data);
es->pindent++;
es->planinfo->append_str_info("%3d --%s\n", plan_node_id, es->planinfo->m_planInfo->m_pname.data);
es->planinfo->set_pname(es->planinfo->m_planInfo->m_pname.data);
if (es->planinfo->m_runtimeinfo)
es->planinfo->m_runtimeinfo->put(
-1, -1, PLAN_NAME, PointerGetDatum(cstring_to_text(es->planinfo->m_planInfo->m_pname.data)));
}
if (es->costs) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty == false) {
appendStringInfo(
es->str, " (cost=%.2f..%.2f rows=%.0f ", plan->startup_cost, plan->total_cost, plan->plan_rows);
#ifndef ENABLE_MULTIPLE_NODES
if (plan->pred_total_time >= 0) {
appendStringInfo(es->str, "p-time=%.0f ", plan->pred_total_time);
}
if (plan->pred_rows >= 0) {
appendStringInfo(es->str, "p-rows=%.0f ", plan->pred_rows);
}
#endif
if ((IsA(plan, HashJoin) || IsA(plan, VecHashJoin)) && es->verbose)
appendStringInfo(es->str,
"distinct=[%.0f, %.0f] ",
plan->outerdistinct > 0 ? ceil(plan->outerdistinct) : 1.0,
plan->innerdistinct > 0 ? ceil(plan->innerdistinct) : 1.0);
appendStringInfo(es->str, "width=%d)", plan->plan_width);
} else {
StringInfoData buf;
initStringInfo(&buf);
appendStringInfo(&buf, "%.2f", plan->total_cost);
es->planinfo->m_planInfo->put(ESTIMATE_COSTS, PointerGetDatum(cstring_to_text(buf.data)));
pfree_ext(buf.data);
#ifndef ENABLE_MULTIPLE_NODES
PredAppendInfo(plan, buf, es);
#endif
es->planinfo->m_planInfo->put(
ESTIMATE_ROWS, DirectFunctionCall1(dround, Float8GetDatum(plan->plan_rows)));
if ((IsA(plan, HashJoin) || IsA(plan, VecHashJoin)) && es->verbose) {
char opdist[130] = "\0";
int rc = EOK;
rc = sprintf_s(opdist,
sizeof(opdist),
"%.0f, %.0f",
plan->outerdistinct > 1 ? ceil(plan->outerdistinct) : 1.0,
plan->innerdistinct > 1 ? ceil(plan->innerdistinct) : 1.0);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ESTIMATE_DISTINCT, PointerGetDatum(cstring_to_text(opdist)));
}
es->planinfo->m_planInfo->put(ESTIMATE_WIDTH, plan->plan_width);
if (es->planinfo->m_planInfo->m_query_mem_mode) {
char memstr[50] = "\0";
if (plan->operatorMemKB[0] > 0) {
int rc = 0;
if (plan->operatorMemKB[0] < 1024)
rc = sprintf_s(memstr, sizeof(memstr), "%dKB", (int)plan->operatorMemKB[0]);
else {
if (plan->operatorMemKB[0] > MIN_OP_MEM && plan->operatorMaxMem > plan->operatorMemKB[0])
rc = sprintf_s(memstr,
sizeof(memstr),
"%dMB(%dMB)",
(int)plan->operatorMemKB[0] / 1024,
(int)plan->operatorMaxMem / 1024);
else
rc = sprintf_s(memstr, sizeof(memstr), "%dMB", (int)plan->operatorMemKB[0] / 1024);
}
securec_check_ss(rc, "\0", "\0");
}
es->planinfo->m_planInfo->put(ESTIMATE_MEMORY, PointerGetDatum(cstring_to_text(memstr)));
}
}
} else {
ExplainPropertyFloat("Startup Cost", plan->startup_cost, 2, es);
ExplainPropertyFloat("Total Cost", plan->total_cost, 2, es);
ExplainPropertyFloat("Plan Rows", plan->plan_rows, 0, es);
ExplainPropertyInteger("Plan Width", plan->plan_width, es);
#ifndef ENABLE_MULTIPLE_NODES
PredGetInfo(plan, es);
#endif
}
}
* We have to forcibly clean up the instrumentation state because we
* haven't done ExecutorEnd yet. This is pretty grotty ...
*/
if (planstate->instrument)
InstrEndLoop(planstate->instrument);
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
show_datanode_time(es, planstate);
show_stream_send_time(es, planstate);
} else {
if (is_pretty) {
if (!es->from_dn) {
show_pretty_time(es, planstate->instrument, NULL, -1, -1, 0);
if (es->planinfo->m_runtimeinfo != NULL) {
if (unlikely(planstate->instrument == NULL)) {
ereport(ERROR,
(errcode(ERRCODE_UNEXPECTED_NULL_VALUE), errmsg("invalid planstate->instrument value.")));
}
get_oper_time<false>(es, planstate, planstate->instrument, -1, -1);
}
}
} else {
show_time<false>(es, planstate->instrument, -1);
}
}
if (es->verbose || es->plan)
show_plan_tlist(planstate, ancestors, es);
if (es->nodes && es->verbose && ng_enable_nodegroup_explain()) {
show_plan_execnodes(planstate, es);
}
switch (nodeTag(plan)) {
case T_NestLoop:
case T_MergeJoin:
case T_AsofJoin:
case T_HashJoin:
if (es->format != EXPLAIN_FORMAT_TEXT || (es->verbose && ((Join *) plan)->inner_unique))
ExplainProperty("Inner Unique", ((Join *) plan)->inner_unique?"true":"false", true, es);
if (is_pretty && es->verbose && ((Join *)plan)->inner_unique) {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str, "Inner Unique: %s\n",
((Join *)plan)->inner_unique ? "true" : "false");
}
break;
default:
break;
}
switch (nodeTag(plan)) {
#ifdef USE_SPQ
case T_SpqIndexScan:
#endif
case T_IndexScan:
show_scan_qual(((IndexScan*)plan)->indexqualorig, "Index Cond", planstate, ancestors, es);
if (((IndexScan*)plan)->indexqualorig)
show_instrumentation_count("Rows Removed by Index Recheck", 2, planstate, es);
show_scan_qual(((IndexScan*)plan)->indexorderbyorig, "Order By", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
break;
#ifdef USE_SPQ
case T_SpqIndexOnlyScan:
#endif
case T_IndexOnlyScan:
show_scan_qual(((IndexOnlyScan*)plan)->indexqual, "Index Cond", planstate, ancestors, es);
if (((IndexOnlyScan*)plan)->indexqual)
show_instrumentation_count("Rows Removed by Index Recheck", 2, planstate, es);
show_scan_qual(((IndexOnlyScan*)plan)->indexorderby, "Order By", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
if (es->analyze)
ExplainPropertyLong("Heap Fetches", ((IndexOnlyScanState*)planstate)->ioss_HeapFetches, es);
break;
case T_AnnIndexScan:
show_scan_qual(((AnnIndexScan*)plan)->indexqualorig, "Index Cond", planstate, ancestors, es);
if (((AnnIndexScan*)plan)->indexqualorig)
show_instrumentation_count("Rows Removed by Index Recheck", 2, planstate, es);
show_scan_qual(((AnnIndexScan*)plan)->indexorderbyorig, "Order By", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
break;
case T_BitmapIndexScan:
show_scan_qual(((BitmapIndexScan*)plan)->indexqualorig, "Index Cond", planstate, ancestors, es);
break;
case T_CStoreIndexCtidScan:
show_scan_qual(((CStoreIndexCtidScan*)plan)->indexqualorig, "Index Cond", planstate, ancestors, es);
break;
case T_CStoreIndexScan:
show_scan_qual(((CStoreIndexScan*)plan)->indexqualorig, "Index Cond", planstate, ancestors, es);
if (((CStoreIndexScan*)plan)->indexqualorig)
show_instrumentation_count("Rows Removed by Index Recheck", 2, planstate, es);
show_scan_qual(((CStoreIndexScan*)plan)->indexorderbyorig, "Order By", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
show_llvm_info(planstate, es);
break;
#ifdef PGXC
case T_ModifyTable:
case T_VecModifyTable: {
ModifyTable* mt = (ModifyTable*)plan;
ListCell* elt = NULL;
ListCell* lc = NULL;
if (mt->operation == CMD_MERGE) {
if (es->analyze) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL
&& es->planinfo->m_detailInfo != NULL) {
es->planinfo->m_detailInfo->set_plan_name<true, false>();
}
show_instrumentation_count("Merge Inserted", 1, planstate, es);
show_instrumentation_count("Merge Updated", 2, planstate, es);
show_instrumentation_count("Merge Deleted", 3, planstate, es);
}
show_modifytable_merge_info(planstate, es);
foreach (lc, mt->mergeActionList) {
MergeAction* action = (MergeAction*)lfirst(lc);
if (action->matched) {
if (mt->remote_update_plans) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "WHEN MATCHED\n");
es->indent++;
foreach (elt, mt->remote_update_plans) {
if (lfirst(elt)) {
ExplainRemoteQuery((RemoteQuery*)lfirst(elt), planstate, ancestors, es);
}
}
es->indent--;
} else
show_scan_qual((List*)action->qual, "Update Cond", planstate, ancestors, es, true);
} else if (!action->matched) {
if (mt->remote_insert_plans) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "WHEN NOT MATCHED\n");
es->indent++;
foreach (elt, mt->remote_insert_plans) {
if (lfirst(elt)) {
ExplainRemoteQuery((RemoteQuery*)lfirst(elt), planstate, ancestors, es);
}
}
es->indent--;
} else
show_scan_qual((List*)action->qual, "Insert Cond", planstate, ancestors, es, true);
}
}
} else {
ModifyTableState* mtstate = (ModifyTableState*)planstate;
if (mtstate->mt_upsert != NULL &&
mtstate->mt_upsert->us_action != UPSERT_NONE && mtstate->resultRelInfo->ri_NumIndices > 0) {
show_on_duplicate_info(mtstate, es, ancestors);
}
foreach (elt, mt->remote_plans) {
if (lfirst(elt)) {
ExplainRemoteQuery((RemoteQuery*)lfirst(elt), planstate, ancestors, es);
}
}
}
} break;
case T_RemoteQuery:
case T_VecRemoteQuery:
show_merge_sort_keys(planstate, ancestors, es);
ExplainRemoteQuery((RemoteQuery*)plan, planstate, ancestors, es);
show_scan_qual(plan->qual, "Coordinator quals", planstate, ancestors, es);
break;
#else
case T_ModifyTable:
show_modifytable_info((ModifyTableState*)planstate, es);
break;
#endif
#ifdef USE_SPQ
case T_SpqBitmapHeapScan:
#endif
case T_BitmapHeapScan:
case T_CStoreIndexHeapScan:
show_scan_qual(((BitmapHeapScan*)plan)->bitmapqualorig, "Recheck Cond", planstate, ancestors, es);
if (((BitmapHeapScan*)plan)->bitmapqualorig)
show_instrumentation_count("Rows Removed by Index Recheck", 2, planstate, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
if (nodeTag(plan) == T_BitmapHeapScan && es->analyze) {
show_tidbitmap_info((BitmapHeapScanState*)planstate, es);
}
show_llvm_info(planstate, es);
break;
case T_StartWithOp:
show_startwith_pseudo_entries(planstate, ancestors, es);
show_startwith_dfx((StartWithOpState*)planstate, es);
break;
case T_SeqScan:
#ifdef USE_SPQ
case T_SpqSeqScan:
#endif
show_tablesample(plan, planstate, ancestors, es);
if (!((SeqScan*)plan)->scanBatchMode) {
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual) {
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
}
}
break;
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
#endif
case T_ValuesScan:
case T_CteScan:
case T_WorkTableScan:
case T_SubqueryScan:
case T_VecSubqueryScan:
show_tablesample(plan, planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
show_llvm_info(planstate, es);
break;
case T_Stream:
case T_VecStream: {
show_merge_sort_keys(planstate, ancestors, es);
ShowStreamRunNodeInfo((Stream*)plan, es);
if (is_pretty && es->planinfo->m_query_summary) {
showStreamnetwork((Stream*)plan, es);
}
} break;
case T_FunctionScan:
if (es->verbose)
show_expression(
((FunctionScan*)plan)->funcexpr, "Function Call", planstate, ancestors, es->verbose, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
break;
case T_TidScan: {
* The tidquals list has OR semantics, so be sure to show it
* as an OR condition.
*/
List* tidquals = ((TidScan*)plan)->tidquals;
if (list_length(tidquals) > 1)
tidquals = list_make1(make_orclause(tidquals));
show_scan_qual(tidquals, "TID Cond", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
} break;
case T_TidRangeScan: {
* The tidrangequals list has AND semantics, so be sure to
* show it as an AND condition.
*/
List *tidquals = ((TidRangeScan *) plan)->tidrangequals;
if (list_length(tidquals) > 1)
tidquals = list_make1(make_andclause(tidquals));
show_scan_qual(tidquals, "TID Cond", planstate, ancestors, es);
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1,
planstate, es);
}
break;
case T_ForeignScan:
case T_VecForeignScan: {
ForeignScan* fScan = (ForeignScan*)plan;
* show_predicate_flag represents the label of non-pushdown
* predicate restriction clause on foreign scan level.
*/
char* show_predicate_flag = NULL;
* show_pushdown_flag represents the label of pushdown predicate
* restriction clause on ORC reader level.
*/
char* show_pushdown_flag = NULL;
if (fScan->scan.scan_qual_optimized) {
show_predicate_flag = FILTER_INFORMATIONAL_CONSTRAINT_FLAG;
} else {
show_predicate_flag = FILTER_FLAG;
}
if (fScan->scan.predicate_pushdown_optimized) {
show_pushdown_flag = PUSHDOWN_FILTER_INFORMATIONAL_CONSTRAINT_FLAG;
} else {
show_pushdown_flag = PUSHDOWN_PREDICATE_FLAG;
}
show_scan_qual(plan->qual, show_predicate_flag, planstate, ancestors, es);
show_pushdown_qual(planstate, ancestors, es, show_pushdown_flag);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
if (es->wlm_statistics_plan_max_digit == NULL) {
show_foreignscan_info((ForeignScanState*)planstate, es);
}
show_storage_filter_info(planstate, es);
show_llvm_info(planstate, es);
if (IsA(plan, VecForeignScan)) {
show_bloomfilter<false>(plan, planstate, ancestors, es);
}
} break;
case T_VecNestLoop:
case T_NestLoop: {
NestLoop* nestLoop = (NestLoop*)plan;
if (nestLoop->join.optimizable) {
show_upper_qual(((NestLoop*)plan)->join.joinqual,
"Join Filter(Informational Constraint Optimization)",
planstate,
ancestors,
es);
} else {
show_upper_qual(((NestLoop*)plan)->join.joinqual, "Join Filter", planstate, ancestors, es);
}
if (((NestLoop*)plan)->join.joinqual)
show_instrumentation_count("Rows Removed by Join Filter", 1, planstate, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 2, planstate, es);
show_llvm_info(planstate, es);
show_skew_optimization(planstate, es);
} break;
case T_VecMergeJoin:
case T_MergeJoin: {
MergeJoin* mergeJoin = (MergeJoin*)plan;
if (mergeJoin->join.optimizable) {
show_upper_qual(((MergeJoin*)plan)->mergeclauses,
"Merge Cond(Informational Constraint Optimization)",
planstate,
ancestors,
es);
} else {
show_upper_qual(((MergeJoin*)plan)->mergeclauses, "Merge Cond", planstate, ancestors, es);
}
show_upper_qual(((MergeJoin*)plan)->join.joinqual, "Join Filter", planstate, ancestors, es);
if (((MergeJoin*)plan)->join.joinqual)
show_instrumentation_count("Rows Removed by Join Filter", 1, planstate, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 2, planstate, es);
show_llvm_info(planstate, es);
show_skew_optimization(planstate, es);
} break;
case T_HashJoin: {
show_upper_qual(((HashJoin*)plan)->hashclauses, "Hash Cond", planstate, ancestors, es);
show_upper_qual(((HashJoin*)plan)->join.joinqual, "Join Filter", planstate, ancestors, es);
if (((HashJoin*)plan)->join.joinqual)
show_instrumentation_count("Rows Removed by Join Filter", 1, planstate, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 2, planstate, es);
show_skew_optimization(planstate, es);
} break;
case T_VecHashJoin: {
show_upper_qual(((HashJoin*)plan)->hashclauses, "Hash Cond", planstate, ancestors, es);
show_upper_qual(((HashJoin*)plan)->join.joinqual, "Join Filter", planstate, ancestors, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
show_llvm_info(planstate, es);
show_bloomfilter<true>(plan, planstate, ancestors, es);
show_skew_optimization(planstate, es);
}
show_vechash_info((VecHashJoinState*)planstate, es);
break;
case T_AsofJoin:
case T_VecAsofJoin: {
show_upper_qual(((VecAsofJoin*)plan)->hashclauses, "Part Cond", planstate, ancestors, es);
show_upper_qual(((VecAsofJoin*)plan)->join.joinqual, "Merge Cond", planstate, ancestors, es);
show_skew_optimization(planstate, es);
} break;
case T_VecAgg:
case T_Agg:
show_groupby_keys((AggState*)planstate, ancestors, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
switch (((Agg*)plan)->aggstrategy) {
case AGG_HASHED: {
show_hashAgg_info((AggState*)planstate, es);
show_llvm_info(planstate, es);
} break;
case AGG_SORTED: {
show_llvm_info(planstate, es);
} break;
default:
break;
}
show_skew_optimization(planstate, es);
show_unique_check_info(planstate, es);
break;
case T_Group:
case T_VecGroup:
show_group_keys((GroupState*)planstate, ancestors, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
break;
case T_Sort:
case T_VecSort:
show_sort_keys((SortState*)planstate, ancestors, es);
show_sort_info((SortState*)planstate, es);
show_llvm_info(planstate, es);
break;
case T_SortGroup: {
SortGroup *plan = (SortGroup *)planstate->plan;
show_sort_group_keys(planstate, "Sorted Group Key", plan->numCols, plan->sortColIdx, plan->sortOperators,
plan->collations, plan->nullsFirst, ancestors, es);
show_sort_group_info(castNode(SortGroupState, planstate), es);
break;
}
case T_MergeAppend:
show_merge_append_keys((MergeAppendState*)planstate, ancestors, es);
break;
case T_BaseResult:
case T_VecResult:
show_upper_qual((List*)((BaseResult*)plan)->resconstantqual, "One-Time Filter", planstate, ancestors, es);
show_upper_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
break;
case T_Hash:
show_hash_info((HashState*)planstate, es);
break;
case T_SetOp:
case T_VecSetOp:
switch (((SetOp*)plan)->strategy) {
case SETOP_HASHED:
show_setop_info((SetOpState*)planstate, es);
break;
default:
break;
}
break;
case T_ExtensiblePlan:
show_pushdown_qual(planstate, ancestors, es, PUSHDOWN_PREDICATE_FLAG);
break;
case T_RecursiveUnion:
show_recursive_info((RecursiveUnionState*)planstate, es);
break;
case T_RowToVec:
if (IsA(plan->lefttree, SeqScan) && ((SeqScan*)plan->lefttree)->scanBatchMode) {
show_scan_qual(plan->lefttree->qual, "Filter", planstate, ancestors, es);
if (plan->lefttree->qual) {
show_instrumentation_count("Rows Removed by Filter", 1, planstate->lefttree, es);
}
}
break;
#ifdef USE_SPQ
case T_AssertOp:
show_upper_qual(plan->qual, "Assert Cond", planstate, ancestors, es);
break;
case T_ShareInputScan:
show_upper_qual(plan->qual, "Shared Scan", planstate, ancestors, es);
break;
case T_Sequence:
show_upper_qual(plan->qual, "Sequence", planstate, ancestors, es);
break;
#endif
default:
break;
}
switch (nodeTag(plan)) {
case T_VecNestLoop:
case T_NestLoop:
case T_VecMergeJoin:
case T_MergeJoin:
case T_Group:
case T_VecGroup:
case T_BaseResult:
case T_VecResult:
case T_Append:
case T_Unique:
case T_VecUnique:
case T_Limit:
case T_PartIterator:
case T_VecPartIterator:
case T_VecToRow:
case T_RowToVec:
case T_VecAppend:
case T_VecLimit:
case T_MergeAppend:
case T_SubqueryScan:
case T_VecSubqueryScan:
case T_ValuesScan:
break;
default: {
if (es->buffers) {
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id))
show_datanode_buffers(es, planstate);
else if (!es->from_dn) {
if (is_pretty && es->planinfo->m_IOInfo)
show_buffers(es, es->planinfo->m_IOInfo->info_str, planstate->instrument, false, -1, -1, NULL);
else
show_buffers(es, es->str, planstate->instrument, false, -1, -1, NULL);
}
}
} break;
}
if (es->cpu) {
uint64 proRows = 0;
double incCycles = 0.0;
double exCycles = 0.0;
double outerCycles = 0.0;
double innerCycles = 0.0;
uint64 outterRows = 0;
uint64 innerRows = 0;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->detail) {
ExplainOpenGroup("Cpus In Detail", "Cpus In Detail", false, es);
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (int j = 0; j < dop; j++) {
outerCycles = 0.0;
innerCycles = 0.0;
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr == NULL)
continue;
append_datanode_name(es, node_name, dop, j);
show_child_cpu_cycles_and_rows<true>(
planstate, i, j, &outerCycles, &innerCycles, &outterRows, &innerRows);
if (es->planinfo != NULL && es->planinfo->m_runtimeinfo != NULL && instr != NULL &&
instr->nloops > 0) {
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
es->planinfo->m_datanodeInfo->set_datanode_name(node_name, j, dop);
}
proRows = (long)(instr->ntuples);
CalculateProcessedRows(planstate, i, j, &innerRows, &outterRows, &proRows);
show_cpu(es, instr, innerCycles, outerCycles, i, j, proRows);
ExplainCloseGroup("Plan", NULL, true, es);
}
}
ExplainCloseGroup("Cpus In Detail", "Cpus In Detail", false, es);
} else
show_detail_cpu(es, planstate);
} else if (!es->from_dn) {
const CPUUsage* usage = &planstate->instrument->cpuusage;
show_child_cpu_cycles_and_rows<false>(planstate, 0, 0, &outerCycles, &innerCycles, &outterRows, &innerRows);
incCycles = usage->m_cycles;
exCycles = incCycles - outerCycles - innerCycles;
proRows = (long)(planstate->instrument->ntuples);
CalculateProcessedRows(planstate, 0, 0, &innerRows, &outterRows, &proRows);
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty) {
int64 ex_cyc_rows = proRows != 0 ? (long)(exCycles / proRows) : 0;
if (es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(-1, -1, EX_CYC, Int64GetDatum(exCycles));
es->planinfo->m_runtimeinfo->put(-1, -1, INC_CYC, Int64GetDatum(incCycles));
es->planinfo->m_runtimeinfo->put(-1, -1, EX_CYC_PER_ROWS, Int64GetDatum(ex_cyc_rows));
appendStringInfoSpaces(es->planinfo->m_datanodeInfo->info_str, 8);
show_cpu_info(es->planinfo->m_datanodeInfo->info_str, incCycles, exCycles, proRows);
}
if (es->planinfo->m_IOInfo) {
es->planinfo->m_IOInfo->set_plan_name<true, true>();
show_cpu_info(es->planinfo->m_IOInfo->info_str, incCycles, exCycles, proRows);
}
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
show_cpu_info(es->str, incCycles, exCycles, proRows);
}
} else {
ExplainPropertyLong("Exclusive Cycles Per Row", proRows != 0 ? (long)(exCycles / proRows) : 0, es);
ExplainPropertyLong("Exclusive Cycles", (long)exCycles, es);
ExplainPropertyLong("Inclusive Cycles", (long)incCycles, es);
}
}
}
switch (nodeTag(plan)) {
case T_SeqScan:
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
#endif
#ifdef USE_SPQ
case T_SpqSeqScan:
case T_SpqIndexScan:
case T_SpqIndexOnlyScan:
#endif
case T_IndexScan:
case T_IndexOnlyScan:
case T_AnnIndexScan:
case T_BitmapHeapScan:
case T_BitmapIndexScan:
case T_CStoreIndexScan:
case T_CStoreIndexHeapScan:
case T_TidScan: {
if (((Scan*)plan)->isPartTbl) {
show_pruning_info(planstate, es, is_pretty);
}
if (es->verbose && ((Scan*)plan)->bucketInfo != NULL) {
show_bucket_info(planstate, es, is_pretty);
}
} break;
case T_PartIterator:
case T_VecPartIterator:
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_pretty == false) {
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
}
ExplainNodePartition(plan, es);
int subPartCnt = 0;
if (GetSubPartitionIterations(plan, es, &subPartCnt)) {
appendStringInfo(es->str, ", Sub Iterations: %d", subPartCnt);
}
appendStringInfoChar(es->str, '\n');
} else {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(
es->planinfo->m_detailInfo->info_str, "Iterations: %d", ((PartIterator*)plan)->itrs);
appendStringInfoChar(es->planinfo->m_detailInfo->info_str, '\n');
}
} else {
ExplainPropertyInteger("Iterations", ((PartIterator*)plan)->itrs, es);
int subPartCnt = 0;
if (GetSubPartitionIterations(plan, es, &subPartCnt)) {
ExplainPropertyInteger("Sub Iterations", subPartCnt, es);
}
}
break;
default:
break;
}
runnext:
haschildren = planstate->initPlan || outerPlanState(planstate) || innerPlanState(planstate) ||
IsA(plan, ModifyTable) || IsA(plan, VecModifyTable) || IsA(plan, Append) || IsA(plan, VecAppend) ||
IsA(plan, MergeAppend) || IsA(plan, VecMergeAppend) || IsA(plan, BitmapAnd) || IsA(plan, BitmapOr) ||
IsA(plan, SubqueryScan) || IsA(plan, VecSubqueryScan) ||
(IsA(planstate, ExtensiblePlanState) && ((ExtensiblePlanState*)planstate)->extensible_ps != NIL) ||
planstate->subPlan;
if (haschildren) {
ExplainOpenGroup("Plans", "Plans", false, es);
ancestors = lcons(planstate, ancestors);
}
if (planstate->initPlan)
ExplainSubPlans(planstate->initPlan, ancestors, "InitPlan", es);
if (IS_STREAM_PLAN && IsA(plan, CteScan)) {
CteScanState* css = (CteScanState*)planstate;
if (css->cteplanstate) {
ExplainNode<is_pretty>(css->cteplanstate, ancestors, "CTE Sub", NULL, es);
}
}
old_dn_flag = es->from_dn;
if (outerPlanState(planstate)) {
if (IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery))
es->from_dn = true;
ExplainNode<is_pretty>(outerPlanState(planstate), ancestors, "Outer", NULL, es);
}
es->from_dn = old_dn_flag;
if (innerPlanState(planstate)) {
if (IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery))
es->from_dn = true;
ExplainNode<is_pretty>(innerPlanState(planstate), ancestors, "Inner", NULL, es);
}
es->from_dn = old_dn_flag;
switch (nodeTag(plan)) {
case T_ModifyTable:
case T_VecModifyTable:
ExplainMemberNodes(((ModifyTable*)plan)->plans, ((ModifyTableState*)planstate)->mt_plans, ancestors, es);
break;
case T_VecAppend:
case T_Append:
ExplainMemberNodes(((Append*)plan)->appendplans, ((AppendState*)planstate)->appendplans, ancestors, es);
break;
case T_MergeAppend:
ExplainMemberNodes(
((MergeAppend*)plan)->mergeplans, ((MergeAppendState*)planstate)->mergeplans, ancestors, es);
break;
case T_BitmapAnd:
ExplainMemberNodes(
((BitmapAnd*)plan)->bitmapplans, ((BitmapAndState*)planstate)->bitmapplans, ancestors, es);
break;
case T_BitmapOr:
ExplainMemberNodes(((BitmapOr*)plan)->bitmapplans, ((BitmapOrState*)planstate)->bitmapplans, ancestors, es);
break;
case T_CStoreIndexAnd:
ExplainMemberNodes(
((CStoreIndexAnd*)plan)->bitmapplans, ((BitmapAndState*)planstate)->bitmapplans, ancestors, es);
break;
case T_CStoreIndexOr:
ExplainMemberNodes(
((CStoreIndexOr*)plan)->bitmapplans, ((BitmapOrState*)planstate)->bitmapplans, ancestors, es);
break;
case T_SubqueryScan:
case T_VecSubqueryScan:
ExplainNode<is_pretty>(((SubqueryScanState*)planstate)->subplan, ancestors, "Subquery", NULL, es);
break;
case T_ExtensiblePlan:
ExplainExtensibleChildren((ExtensiblePlanState*)planstate, ancestors, es);
break;
#ifdef USE_SPQ
case T_Sequence:
ExplainMemberNodes(((Sequence*)plan)->subplans, ((SequenceState*)planstate)->subplans, ancestors, es);
break;
#endif
default:
break;
}
if (planstate->subPlan)
ExplainSubPlans(planstate->subPlan, ancestors, "SubPlan", es);
if (haschildren) {
ancestors = list_delete_first(ancestors);
ExplainCloseGroup("Plans", "Plans", false, es);
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
es->indent = save_indent;
es->pindent = save_pindent;
}
ExplainCloseGroup("Plan", relationship ? NULL : "Plan", true, es);
* update". */
if (es->plan && (planstate->plan->plan_node_id != 0 || IsExplainPlanSelectForUpdateStmt)) {
* 1.Handle the case for select for update.
* Step 1: Check if it's a select for update case.
*/
if (planstate->plan->plan_node_id == 0 && IsExplainPlanSelectForUpdateStmt) {
GetPlanNodePlainText(plan, &pname, &sname, &strategy, &operation, &pt_operation, &pt_options);
RangeTblEntry* rte = (RangeTblEntry*)llast(es->rtable);
char* objectname = rte->relname;
const char* object_type = "REMOTE_QUERY";
es->planinfo->m_planTableData->set_plan_table_objs(plan->plan_node_id, objectname, object_type, NULL);
show_plan_tlist(planstate, ancestors, es);
}
if (pt_operation == NULL && pname != NULL)
es->planinfo->m_planTableData->set_plan_table_streaming_ops(pname, &pt_operation, &pt_options);
es->planinfo->m_planTableData->set_plan_table_ops(planstate->plan->plan_node_id, pt_operation, pt_options);
* 3.set object info for index.
* For index scan we cannot get its object info from ExplainTargetRel. We only can get the info from
* ExplainNode.
*/
if (strcmp(pt_operation, "INDEX") == 0 && pt_index_name != NULL && pt_index_owner != NULL)
es->planinfo->m_planTableData->set_plan_table_objs(
planstate->plan->plan_node_id, pt_index_name, pt_operation, pt_index_owner);
es->planinfo->m_planTableData->set_plan_table_cost_card(plan->plan_node_id, plan->total_cost, plan->plan_rows);
}
#ifdef USE_SPQ
es->current_id = current_id_bak;
#endif
}
* CalculateProcessedRows
* The input processed tuple number equals to the returned tuple number
* of the current node. Here we modify the iput processed tuple rows more precisely.
* For Example in SeqScan, the processed tuple number should add the tuple number
* removed by filter.
* @param (in) innerRows:
* returned tuple number for right child node.
* @param (in) outterRows:
* returned tuple number for left child node.
* @param (in) processedrows:
* returned tuple number for current node.
* @param (out) processedrows:
* the processed tuple number for the current node.
*/
static void CalculateProcessedRows(
PlanState* planstate, int idx, int smpIdx, const uint64* inner_rows, const uint64* outter_rows, uint64* processed_rows)
{
int removed_rows = 0;
Plan* plan = planstate->plan;
switch (nodeTag(plan)) {
case T_SeqScan:
if (!((SeqScan*)plan)->scanBatchMode) {
show_removed_rows(1, planstate, idx, smpIdx, &removed_rows);
*processed_rows += removed_rows;
}
break;
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
#endif
case T_NestLoop:
case T_VecNestLoop:
case T_VecMergeJoin:
case T_MergeJoin:
case T_AsofJoin:
case T_VecAsofJoin:
case T_HashJoin:
case T_VecHashJoin:
*processed_rows = (*inner_rows) + (*outter_rows);
break;
case T_Agg:
case T_Sort:
case T_SetOp:
case T_VecSetOp:
case T_VecAgg:
case T_VecSort:
*processed_rows = *outter_rows;
break;
default:
break;
}
}
* show_plan_execnodes
* show exec nodes information of a plan node
*
* @param (in) planstate:
* the plan state
* @param (in) es:
* the explain state
*
* @return: void
*/
static void show_plan_execnodes(PlanState* planstate, ExplainState* es)
{
Plan* plan = planstate->plan;
if (IsA(plan, Stream) || IsA(plan, VecStream) || IsA(plan, ModifyTable) || IsA(plan, VecModifyTable) ||
IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery) || u_sess->pgxc_cxt.NumDataNodes == 0) {
return;
}
ShowRunNodeInfo(plan->exec_nodes, es, "Exec Nodes");
return;
}
* Show the targetlist of a plan node
*/
static void show_plan_tlist(PlanState* planstate, List* ancestors, ExplainState* es)
{
Plan* plan = planstate->plan;
List* context = NIL;
List* result = NIL;
bool useprefix = false;
ListCell* lc = NULL;
if (plan->targetlist == NIL)
return;
if (IsA(plan, Append) || IsA(plan, VecAppend))
return;
if (IsA(plan, MergeAppend) || IsA(plan, VecMergeAppend))
return;
if (IsA(plan, RecursiveUnion) || IsA(plan, VecRecursiveUnion))
return;
context = deparse_context_for_planstate((Node*)planstate, ancestors, es->rtable);
useprefix = list_length(es->rtable) > 1;
foreach (lc, plan->targetlist) {
TargetEntry* tle = (TargetEntry*)lfirst(lc);
if (IsPseudoReturnTargetEntry(tle) || IsPseudoInternalTargetEntry(tle)) {
continue;
}
result = lappend(result, deparse_expression((Node*)tle->expr, context, useprefix, false));
}
if (IsA(plan, ForeignScan) || IsA(plan, VecForeignScan)) {
ForeignScan* fscan = (ForeignScan*)plan;
if (OidIsValid(fscan->scan_relid) && IsSpecifiedFDWFromRelid(fscan->scan_relid, GC_FDW)) {
List* str_targetlist = get_str_targetlist(fscan->fdw_private);
if (str_targetlist != NULL)
result = str_targetlist;
}
}
ExplainPropertyList("Output", result, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "Output: ");
ExplainPrettyList(result, es);
}
if (es->plan)
es->planinfo->m_planTableData->set_plan_table_projection(plan->plan_node_id, result);
#ifdef ENABLE_MULTIPLE_NODES
if (plan->distributed_keys != NIL && show_scan_distributekey(plan)) {
List* distKey = NIL;
List* keyList = NIL;
keyList = plan->distributed_keys;
foreach (lc, keyList) {
Var* distriVar = (Var*)lfirst(lc);
distKey = lappend(distKey, deparse_expression((Node*)distriVar, context, useprefix, false));
}
ExplainPropertyList("Distribute Key", distKey, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfoString(es->planinfo->m_verboseInfo->info_str, "Distribute Key: ");
ExplainPrettyList(distKey, es);
}
}
#endif
#ifdef STREAMPLAN
if ((IsA(plan, Stream) || IsA(plan, VecStream)) &&
(is_redistribute_stream((Stream*)plan) || is_hybid_stream((Stream*)plan)) &&
((Stream*)plan)->distribute_keys != NIL) {
List* distKey = NIL;
List* keyList = NIL;
keyList = ((Stream*)plan)->distribute_keys;
foreach (lc, keyList) {
Var* distriVar = (Var*)lfirst(lc);
distKey = lappend(distKey, deparse_expression((Node*)distriVar, context, useprefix, false));
}
ExplainPropertyList("Distribute Key", distKey, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfoString(es->planinfo->m_verboseInfo->info_str, "Distribute Key: ");
ExplainPrettyList(distKey, es);
}
list_free_deep(distKey);
}
#endif
}
* ContainRteInRlsPolicyWalker
* Check node whether contain specified RangeTblEntry
*
* @param (in)node: Node info
* @param (in)unused_info: do not use this infor
* @return: Find row level security policy in this RangeTblEntry or not
*/
static bool ContainRlsQualInRteWalker(Node* node, void* unused_info)
{
if (node == NULL) {
return false;
}
if (IsA(node, RangeTblEntry)) {
RangeTblEntry* rte = (RangeTblEntry*)node;
if ((rte->rtekind == RTE_RELATION) && (rte->securityQuals != NIL)) {
return true;
}
return false;
}
if (IsA(node, Query)) {
return range_table_walker(((Query*)node)->rtable,
(bool (*)())ContainRlsQualInRteWalker,
unused_info,
QTW_EXAMINE_RTES | QTW_IGNORE_DUMMY);
}
return expression_tree_walker(node, (bool (*)())ContainRlsQualInRteWalker, unused_info);
}
* Show a generic expression
*/
static void show_expression(
Node* node, const char* qlabel, PlanState* planstate, List* ancestors, bool useprefix, ExplainState* es)
{
List* context = NIL;
char* exprstr = NULL;
context = deparse_context_for_planstate((Node*)planstate, ancestors, es->rtable);
exprstr = deparse_expression(node, context, useprefix, false);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_detailInfo) {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str, "%s: %s\n", qlabel, exprstr);
}
ExplainPropertyText(qlabel, exprstr, es);
}
* Show a qualifier expression (which is a List with implicit AND semantics)
*/
static void show_qual(
List* qual, const char* qlabel, PlanState* planstate, List* ancestors, bool useprefix, ExplainState* es)
{
Node* node = NULL;
if (qual == NIL)
return;
node = (Node*)make_ands_explicit(qual);
show_expression(node, qlabel, planstate, ancestors, useprefix, es);
}
* Show a qualifier expression for a scan plan node
*/
static void show_scan_qual(List *qual, const char *qlabel, PlanState *planstate, List *ancestors, ExplainState *es,
bool show_prefix)
{
bool useprefix = false;
useprefix =
(show_prefix || IsA(planstate->plan, SubqueryScan) || IsA(planstate->plan, VecSubqueryScan) || es->verbose);
show_qual(qual, qlabel, planstate, ancestors, useprefix, es);
}
* @Description: Show bloom filter index info.
* @in filterNumList: Index list.
* @in es: Explain state.
*/
template <bool generate>
static void show_bloomfilter_number(const List* filterNumList, ExplainState* es)
{
if (filterNumList != NULL) {
StringInfo str = NULL;
StringInfoData index_info;
initStringInfo(&index_info);
bool first = true;
ListCell* l = NULL;
foreach (l, filterNumList) {
if (first) {
appendStringInfo(&index_info, "%d", lfirst_int(l));
} else {
appendStringInfo(&index_info, ", %d", lfirst_int(l));
}
first = false;
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
es->planinfo->m_detailInfo->set_plan_name<false, true>();
str = es->planinfo->m_detailInfo->info_str;
} else {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoSpaces(es->str, es->indent * 2);
str = es->str;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (generate) {
appendStringInfo(str, "%s: %s\n", "Generate Bloom Filter On Index", index_info.data);
} else {
appendStringInfo(str, "%s: %s\n", "Filter By Bloom Filter On Index", index_info.data);
}
} else {
if (generate) {
ExplainPropertyText("Generate Bloom Filter On Index", index_info.data, es);
} else {
ExplainPropertyText("Filter By Bloom Filter On Index", index_info.data, es);
}
}
pfree_ext(index_info.data);
}
}
* @Description: Show bloomfilter information, include filter var and filter index.
* @in plan: Vechashjoin plan or Scan plan(include hdfs, hdfs foreign scan).
* @in planstate: PlanState node.
* @in ancestors: Ancestors list should already contain the immediate parent of these
* SubPlanStates.
* @in es: Explain state.
*/
template <bool generate>
static void show_bloomfilter(Plan* plan, PlanState* planstate, List* ancestors, ExplainState* es)
{
if (plan->var_list) {
if (generate) {
show_expression((Node*)plan->var_list, "Generate Bloom Filter On Expr", planstate, ancestors, true, es);
} else {
show_expression((Node*)plan->var_list, "Filter By Bloom Filter On Expr", planstate, ancestors, true, es);
}
show_bloomfilter_number<generate>(plan->filterIndexList, es);
}
}
static void show_skew_optimization(const PlanState* planstate, ExplainState* es)
{
#ifdef ENABLE_MULTIPLE_NODES
int skew_opt = SKEW_RES_NONE;
char* skew_txt = NULL;
switch (planstate->plan->type) {
case T_NestLoop:
case T_VecNestLoop:
case T_HashJoin:
case T_VecHashJoin:
case T_AsofJoin:
case T_VecAsofJoin:
case T_MergeJoin:
case T_VecMergeJoin:
skew_opt = ((Join*)planstate->plan)->skewoptimize;
skew_txt = "Skew Join Optimized";
break;
case T_Agg:
case T_VecAgg:
skew_opt = ((Agg*)planstate->plan)->skew_optimize;
skew_txt = "Skew Agg Optimized";
break;
default:
return;
}
if (skew_opt == SKEW_RES_NONE)
return;
StringInfoData str;
initStringInfo(&str);
appendStringInfo(&str, "%s by", skew_txt);
if ((unsigned int)skew_opt & SKEW_RES_HINT) {
if ((unsigned int)skew_opt & (SKEW_RES_HINT - 1))
appendStringInfo(&str, " %s,", "Hint");
else
appendStringInfo(&str, " %s", "Hint");
}
if ((unsigned int)skew_opt & SKEW_RES_RULE) {
if ((unsigned int)skew_opt & (SKEW_RES_RULE - 1))
appendStringInfo(&str, " %s,", "Rule");
else
appendStringInfo(&str, " %s", "Rule");
}
if ((unsigned int)skew_opt & SKEW_RES_STAT)
appendStringInfo(&str, " %s", "Statistic");
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
if (planstate->plan->type == T_Agg || planstate->plan->type == T_VecAgg) {
StringInfo agg_node_des = makeStringInfo();
int rc = snprintf_s(agg_node_des->data,
agg_node_des->maxlen,
agg_node_des->maxlen - 1,
"%3d --%s",
planstate->plan->plan_node_id,
es->planinfo->m_planInfo->m_pname.data);
securec_check_ss(rc, "\0", "\0");
if (strstr(es->planinfo->m_detailInfo->info_str->data, agg_node_des->data) == NULL)
appendStringInfo(es->planinfo->m_detailInfo->info_str,
"%3d --%s\n",
planstate->plan->plan_node_id,
es->planinfo->m_planInfo->m_pname.data);
}
appendStringInfoSpaces(es->planinfo->m_detailInfo->info_str, 8);
appendStringInfo(es->planinfo->m_detailInfo->info_str, "%s\n", str.data);
} else {
if (es->format == EXPLAIN_FORMAT_TEXT && !es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "%s\n", str.data);
}
}
pfree_ext(str.data);
#else
return;
#endif
}
* @Description: Show pushdown quals in the flag identifier.
*/
static void show_pushdown_qual(PlanState* planstate, List* ancestors, ExplainState* es, const char* identifier)
{
Plan* plan = planstate->plan;
switch (nodeTag(plan)) {
case T_ForeignScan:
case T_VecForeignScan: {
* Only the HDFS foreign plan has pushdown predicate feature,
* So show the pushdown predicate filter on HDFS foreign plan.
*/
if (((ForeignScanState*)planstate)->fdwroutine->GetFdwType != NULL &&
((ForeignScanState*)planstate)->fdwroutine->GetFdwType() == HDFS_ORC) {
DefElem* private_data = (DefElem*)linitial(((ForeignScan*)plan)->fdw_private);
DfsPrivateItem* item = (DfsPrivateItem*)private_data->arg;
show_scan_qual(item->hdfsQual, identifier, planstate, ancestors, es);
}
break;
}
case T_ExtensiblePlan: {
ExtensiblePlanState* css = (ExtensiblePlanState*)planstate;
show_scan_qual(plan->qual, "Filter", planstate, ancestors, es);
if (plan->qual) {
show_instrumentation_count("Rows Removed by Filter", 1, planstate, es);
}
if (css->methods->ExplainExtensiblePlan) {
css->methods->ExplainExtensiblePlan(css, ancestors, es);
}
break;
}
default: {
* Do nothing.
*/
break;
}
}
}
* Show a qualifier expression for an upper-level plan node
*/
static void show_upper_qual(List* qual, const char* qlabel, PlanState* planstate, List* ancestors, ExplainState* es)
{
bool useprefix = false;
useprefix = (list_length(es->rtable) > 1 || es->verbose);
show_qual(qual, qlabel, planstate, ancestors, useprefix, es);
}
* @Description: append the datanode name to str.
* @in es: explain state.
* @in node_name: datanode name
* @in dop: query dop
* @in j: current dop index
* @out: return void
*/
static void append_datanode_name(ExplainState* es, char* node_name, int dop, int j)
{
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfoSpaces(es->str, es->indent * 2);
if (dop == 1)
appendStringInfo(es->str, " %s ", node_name);
else
appendStringInfo(es->str, " %s[worker %d] ", node_name, j);
} else {
if (dop == 1) {
ExplainPropertyText("DN Name", node_name, es);
} else {
ExplainPropertyText("DN Name", node_name, es);
ExplainPropertyInteger("Worker", j, es);
}
}
}
* Show the group by keys for a Agg node.
*/
static void show_groupby_keys(AggState* aggstate, List* ancestors, ExplainState* es)
{
Agg* plan = (Agg*)aggstate->ss.ps.plan;
if (plan->numCols > 0 || plan->groupingSets) {
ancestors = lcons(aggstate, ancestors);
if (plan->groupingSets)
show_grouping_sets(outerPlanState(aggstate), plan, ancestors, es);
else
show_sort_group_keys(outerPlanState(aggstate),
"Group By Key",
plan->numCols,
plan->grpColIdx,
NULL,
NULL,
NULL,
ancestors,
es);
ancestors = list_delete_first(ancestors);
}
}
static void show_grouping_sets(PlanState* planstate, Agg* agg, List* ancestors, ExplainState* es)
{
List* context = NIL;
bool useprefix = false;
ListCell* lc = NULL;
context = deparse_context_for_planstate((Node*)planstate, ancestors, es->rtable);
useprefix = (list_length(es->rtable) > 1 || es->verbose);
ExplainOpenGroup("Grouping Sets", "Grouping Sets", false, es);
show_grouping_set_keys(planstate, agg, NULL, context, useprefix, ancestors, es);
foreach (lc, agg->chain) {
Agg* aggnode = (Agg*)lfirst(lc);
Sort* sortnode = (Sort*)aggnode->plan.lefttree;
show_grouping_set_keys(planstate, aggnode, sortnode, context, useprefix, ancestors, es);
}
ExplainCloseGroup("Grouping Sets", "Grouping Sets", false, es);
}
static void show_grouping_set_keys(PlanState* planstate, Agg* aggnode, Sort* sortnode, List* context, bool useprefix,
List* ancestors, ExplainState* es)
{
Plan* plan = planstate->plan;
char* exprstr = NULL;
ListCell* lc = NULL;
List* gsets = aggnode->groupingSets;
AttrNumber* keycols = aggnode->grpColIdx;
ExplainOpenGroup("Grouping Set", NULL, true, es);
if (sortnode != NULL) {
show_sort_group_keys(planstate,
"Sort Key",
sortnode->numCols,
sortnode->sortColIdx,
sortnode->sortOperators,
sortnode->collations,
sortnode->nullsFirst,
ancestors,
es);
if (es->format == EXPLAIN_FORMAT_TEXT)
es->indent++;
}
ExplainOpenGroup("Group Keys", "Group Keys", false, es);
foreach (lc, gsets) {
List* result = NIL;
ListCell* lc2 = NULL;
foreach (lc2, (List*)lfirst(lc)) {
Index i = lfirst_int(lc2);
AttrNumber keyresno = keycols[i];
TargetEntry* target = get_tle_by_resno(plan->targetlist, keyresno);
if (target == NULL)
ereport(ERROR,
(errcode(ERRCODE_OPTIMIZER_INCONSISTENT_STATE), errmsg("no tlist entry for key %d", keyresno)));
exprstr = deparse_expression((Node*)target->expr, context, useprefix, true);
result = lappend(result, exprstr);
}
if (result == NIL && es->format == EXPLAIN_FORMAT_TEXT)
ExplainPropertyText("Group By Key", "()", es);
else
ExplainPropertyListNested("Group By Key", result, es);
}
ExplainCloseGroup("Group Keys", "Group Keys", false, es);
if (sortnode != NULL && es->format == EXPLAIN_FORMAT_TEXT)
es->indent--;
ExplainCloseGroup("Grouping Set", NULL, true, es);
}
* @Description:Show the grouping keys for a Group node.
* @in gstate - group node state
* @in ancestors - a list of parent PlanState nodes, most-closely-nested first.
* These are needed in order to interpret PARAM_EXEC Params.
* @in es - explain state.
*/
static void show_group_keys(GroupState* gstate, List* ancestors, ExplainState* es)
{
Group* plan = (Group*)gstate->ss.ps.plan;
ancestors = lcons(gstate, ancestors);
show_sort_group_keys(
outerPlanState(gstate), "Group By Key", plan->numCols, plan->grpColIdx, NULL, NULL, NULL, ancestors, es);
ancestors = list_delete_first(ancestors);
}
* Show the sort keys for a Sort node.
*/
static void show_sort_keys(SortState* sortstate, List* ancestors, ExplainState* es)
{
Sort* plan = (Sort*)sortstate->ss.ps.plan;
show_sort_group_keys((PlanState*)sortstate,
"Sort Key",
plan->numCols,
plan->sortColIdx,
plan->sortOperators,
plan->collations,
plan->nullsFirst,
ancestors,
es);
}
* Likewise, for a MergeAppend node.
*/
static void show_merge_append_keys(MergeAppendState* mstate, List* ancestors, ExplainState* es)
{
MergeAppend* plan = (MergeAppend*)mstate->ps.plan;
show_sort_group_keys((PlanState*)mstate,
"Sort Key",
plan->numCols,
plan->sortColIdx,
plan->sortOperators,
plan->collations,
plan->nullsFirst,
ancestors,
es);
}
* Likewise, for a merge sort info of RemoteQuery or Broadcast node.
*/
static void show_merge_sort_keys(PlanState* state, List* ancestors, ExplainState* es)
{
SimpleSort* sort = NULL;
Plan* plan = state->plan;
if (IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery))
sort = ((RemoteQuery*)plan)->sort;
else if (IsA(plan, Stream) || IsA(plan, VecStream))
sort = ((Stream*)plan)->sort;
if (sort == NULL)
return;
show_sort_group_keys((PlanState*)state,
"Merge Sort Key",
sort->numCols,
sort->sortColIdx,
sort->sortOperators,
sort->sortCollations,
sort->nullsFirst,
ancestors,
es);
}
static void show_startwith_pseudo_entries(PlanState* state, List* ancestors, ExplainState* es)
{
Plan* plan = state->plan;
StartWithOp *swplan = (StartWithOp *)plan;
List* result = NIL;
const char *qlabel = "Start With pseudo atts";
ListCell *lc = NULL;
foreach (lc, swplan->internalEntryList) {
TargetEntry *entry = (TargetEntry *)lfirst(lc);
result = lappend(result, pstrdup(entry->resname));
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "%s: ", qlabel);
ExplainPrettyList(result, es);
} else {
ExplainPropertyList(qlabel, result, es);
}
list_free_deep(result);
}
static void show_detail_sortinfo(ExplainState* es, const char* sortMethod, const char* spaceType, long spaceUsed)
{
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, "Sort Method: %s %s: %ldkB\n", sortMethod, spaceType, spaceUsed);
} else {
ExplainPropertyText("Sort Method", sortMethod, es);
ExplainPropertyLong("Sort Space Used", spaceUsed, es);
ExplainPropertyText("Sort Space Type", spaceType, es);
}
}
* show min and max sort info
*/
static void show_detail_sortinfo_min_max(
ExplainState* es, const char* sortMethod, const char* spaceType, long minspaceUsed, long maxspaceUsed)
{
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(
es->str, "Sort Method: %s %s: %ldkB ~ %ldkB\n", sortMethod, spaceType, minspaceUsed, maxspaceUsed);
} else {
ExplainPropertyText("Sort Method", sortMethod, es);
ExplainPropertyLong("Sort Space Min Used", minspaceUsed, es);
ExplainPropertyLong("Sort Space Max Used", maxspaceUsed, es);
ExplainPropertyText("Sort Space Type", spaceType, es);
}
}
* @Description: show peak memory for each datanode and coordinator
* @in es - the explain state info
* @in plan_size - current plan size
* @out - void
*/
static void show_peak_memory(ExplainState* es, int plan_size)
{
if (u_sess->instr_cxt.global_instr == NULL) {
elog(ERROR, "u_sess->instr_cxt.global_instr is NULL");
}
bool from_datanode = false;
bool last_datanode = u_sess->instr_cxt.global_instr->isFromDataNode(1);
for (int i = 1; i < plan_size; i++) {
int dop = u_sess->instr_cxt.global_instr->getNodeDop(i + 1);
from_datanode = u_sess->instr_cxt.global_instr->isFromDataNode(i + 1);
if (!last_datanode && from_datanode) {
int64 peak_memory = (int64)(unsigned)((uint)(t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->peakChunksQuery -
t_thrd.shemem_ptr_cxt.mySessionMemoryEntry->initMemInChunks)
<< (unsigned int)(chunkSizeInBits - BITS_IN_MB));
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Coordinator Query Peak Memory:\n");
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Query Peak Memory: %ldMB\n", peak_memory);
Instrumentation* instr = NULL;
int64 max_peak_memory = -1;
int64 min_peak_memory = LONG_MAX;
bool is_execute = false;
if (es->detail)
appendStringInfo(es->planinfo->m_staticInfo->info_str, "DataNode Query Peak Memory\n");
for (int j = 0; j < u_sess->instr_cxt.global_instr->getInstruNodeNum(); j++) {
for (int k = 0; k < dop; k++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(j, i + 1, k);
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(j, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", j);
}
#endif
if (instr != NULL) {
if (!is_execute)
is_execute = true;
if (es->detail) {
peak_memory = instr->memoryinfo.peakNodeMemory;
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Query Peak Memory: %ldMB\n", peak_memory);
} else {
max_peak_memory = rtl::max(max_peak_memory, instr->memoryinfo.peakNodeMemory);
min_peak_memory = rtl::min(min_peak_memory, instr->memoryinfo.peakNodeMemory);
}
}
}
}
if (is_execute && !es->detail) {
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Datanode:\n");
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Max Query Peak Memory: %ldMB\n", max_peak_memory);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Min Query Peak Memory: %ldMB\n", min_peak_memory);
}
}
last_datanode = from_datanode;
}
}
* @Description: show the min/max of datanode executor start/run/end time.
* @in es - the explain state info
* @in plan_node_id - current plan node id
* @in stage - the stage of executor time
* @out - void
*/
static void show_dn_executor_time(ExplainState* es, int plan_node_id, ExecutorTime stage)
{
Instrumentation* instr = NULL;
double min_time = 0;
double max_time = 0;
int min_idx = 0;
int max_idx = 0;
bool fisrt_time = true;
const char* symbol_time = "nostage";
int data_size = u_sess->instr_cxt.global_instr->getInstruNodeNum();
for (int i = 0; i < data_size; i++) {
double exec_time;
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, plan_node_id + 1);
if (instr != NULL) {
switch (stage) {
case DN_START_TIME:
exec_time = instr->init_time * 1000.0;
symbol_time = "start";
break;
case DN_RUN_TIME:
exec_time = instr->run_time * 1000.0;
symbol_time = "run";
break;
case DN_END_TIME:
exec_time = instr->end_time * 1000.0;
symbol_time = "end";
break;
}
if (fisrt_time) {
min_time = exec_time;
min_idx = i;
max_time = exec_time;
max_idx = i;
fisrt_time = false;
} else {
if (exec_time > max_time) {
max_time = exec_time;
max_idx = i;
}
if (exec_time < min_time) {
min_time = exec_time;
min_idx = i;
}
}
}
}
#ifdef ENABLE_MULTIPLE_NODES
char* min_node_name = PGXCNodeGetNodeNameFromId(min_idx, PGXC_NODE_DATANODE);
char* max_node_name = PGXCNodeGetNodeNameFromId(max_idx, PGXC_NODE_DATANODE);
#else
char* min_node_name = NULL;
char* max_node_name = NULL;
if (!IS_SPQ_RUNNING) {
min_node_name = g_instance.exec_cxt.nodeName;
max_node_name = g_instance.exec_cxt.nodeName;
} else {
min_node_name = (char*)palloc0(SPQNODENAMELEN);
max_node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(min_node_name, "%d", min_idx);
sprintf(max_node_name, "%d", max_idx);
}
#endif
appendStringInfo(es->planinfo->m_query_summary->info_str, "Datanode executor %s time [%s, %s]: [%.3f ms,%.3f ms]\n", symbol_time, min_node_name, max_node_name, min_time, max_time);
}
* If it's EXPLAIN ANALYZE, show tuplesort stats for a sort node
*/
static void show_sort_info(SortState* sortstate, ExplainState* es)
{
PlanState* planstate = NULL;
planstate = (PlanState*)sortstate;
AssertEreport(IsA(sortstate, SortState) || IsA(sortstate, VecSortState), MOD_EXECUTOR, "unexpect sortstate type");
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
const char* sortMethod = NULL;
const char* spaceType = NULL;
long spaceUsed = 0;
bool has_sort_info = false;
long max_diskUsed = 0;
long min_diskUsed = MIN_DISK_USED;
long max_memoryUsed = 0;
long min_memoryUsed = u_sess->attr.attr_memory.work_mem;
int sortMethodId = 0;
int spaceTypeId = 0;
if (es->detail) {
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
* when this operator does not actual executed, we will not show sort info.
*/
if (instr == NULL || instr->sorthashinfo.sortMethodId < (int)HEAPSORT ||
instr->sorthashinfo.sortMethodId > (int)STILLINPROGRESS)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
sortMethodId = instr->sorthashinfo.sortMethodId;
spaceTypeId = instr->sorthashinfo.spaceTypeId;
sortMethod = sortmessage[sortMethodId].sortName;
spaceUsed = instr->sorthashinfo.spaceUsed;
if (spaceTypeId == SORT_IN_DISK)
spaceType = "Disk";
else
spaceType = "Memory";
if (es->planinfo != NULL && es->planinfo->m_runtimeinfo != NULL) {
es->planinfo->m_runtimeinfo->put(i, 0, SORT_METHOD, PointerGetDatum(cstring_to_text(sortMethod)));
es->planinfo->m_runtimeinfo->put(i, 0, SORT_TYPE, PointerGetDatum(cstring_to_text(spaceType)));
es->planinfo->m_runtimeinfo->put(i, 0, SORT_SPACE, Int64GetDatum(spaceUsed));
if (es->planinfo->m_staticInfo != NULL) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Sort Method: %s %s: %ldkB\n",
sortMethod,
spaceType,
spaceUsed);
}
continue;
}
if (has_sort_info == false)
ExplainOpenGroup("Sort Detail", "Sort Detail", false, es);
has_sort_info = true;
ExplainOpenGroup("Plan", NULL, true, es);
append_datanode_name(es, node_name, 1, 0);
show_detail_sortinfo(es, sortMethod, spaceType, spaceUsed);
ExplainCloseGroup("Plan", NULL, true, es);
}
if (has_sort_info)
ExplainCloseGroup("Sort Detail", "Sort Detail", false, es);
} else {
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL || instr->sorthashinfo.sortMethodId < (int)HEAPSORT ||
instr->sorthashinfo.sortMethodId > (int)STILLINPROGRESS)
continue;
sortMethodId = instr->sorthashinfo.sortMethodId;
spaceTypeId = instr->sorthashinfo.spaceTypeId;
sortMethod = sortmessage[sortMethodId].sortName;
spaceUsed = instr->sorthashinfo.spaceUsed;
if (spaceTypeId == SORT_IN_DISK) {
spaceType = "Disk";
max_diskUsed = rtl::max(spaceUsed, max_diskUsed);
min_diskUsed = rtl::min(spaceUsed, min_diskUsed);
if (min_diskUsed == MIN_DISK_USED)
min_diskUsed = spaceUsed;
} else {
spaceType = "Memory";
max_memoryUsed = rtl::max(spaceUsed, max_memoryUsed);
min_memoryUsed = rtl::min(spaceUsed, min_memoryUsed);
}
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL
&& es->planinfo->m_staticInfo != NULL) {
if (max_memoryUsed != 0) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
}
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Sort Method: %s %s: %ldkB ~ %ldkB\n",
sortMethod,
"Memory",
min_memoryUsed,
max_memoryUsed);
if (max_diskUsed != 0) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
}
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Sort Method: %s %s: %ldkB ~ %ldkB\n",
sortMethod,
"Disk",
min_diskUsed,
max_diskUsed);
} else {
if (max_memoryUsed != 0) {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoSpaces(es->str, es->indent * 2);
show_detail_sortinfo_min_max(es, sortMethod, "Memory", min_memoryUsed, max_memoryUsed);
}
if (max_diskUsed != 0) {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoSpaces(es->str, es->indent * 2);
show_detail_sortinfo_min_max(es, sortMethod, "Disk", min_diskUsed, max_diskUsed);
}
}
}
} else {
char* sortMethod = NULL;
char* spaceType = NULL;
int spaceTypeId = 0;
if (es->analyze && sortstate->sort_Done && sortstate->sortMethodId >= (int)HEAPSORT &&
sortstate->sortMethodId <= (int)STILLINPROGRESS &&
(sortstate->spaceTypeId == SORT_IN_DISK || sortstate->spaceTypeId == SORT_IN_MEMORY)) {
sortMethod = sortmessage[sortstate->sortMethodId].sortName;
spaceTypeId = sortstate->spaceTypeId;
if (spaceTypeId == SORT_IN_DISK)
spaceType = "Disk";
else
spaceType = "Memory";
if (es->planinfo && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(-1, -1, SORT_METHOD, PointerGetDatum(cstring_to_text(sortMethod)));
es->planinfo->m_runtimeinfo->put(-1, -1, SORT_TYPE, PointerGetDatum(cstring_to_text(spaceType)));
es->planinfo->m_runtimeinfo->put(-1, -1, SORT_SPACE, Int64GetDatum(sortstate->spaceUsed));
}
if (es->detail == false && t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo &&
es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Sort Method: %s %s: %ldkB\n",
sortMethod,
spaceType,
sortstate->spaceUsed);
} else {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoSpaces(es->str, es->indent * 2);
show_detail_sortinfo(es, sortMethod, spaceType, sortstate->spaceUsed);
}
}
}
}
* If it's EXPLAIN ANALYZE, show stats for a SortGroupState
*/
static void show_sort_group_info(SortGroupState *state, ExplainState *es)
{
if (!es->analyze)
return;
if (state->sort_Done && state->state != NULL) {
int64 spaceUsed = Max(1, state->spaceUsed / 1024);
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (es->str->len == 0 || es->str->data[es->str->len - 1] == '\n')
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Space Used: %s : " INT64_FORMAT "kB\n", state->spaceType,
spaceUsed);
}
else {
ExplainPropertyInteger("Sort Space Used(kB)", spaceUsed, es);
ExplainPropertyText("Sort Space Type", state->spaceType, es);
}
}
}
template <bool is_detail>
static void show_datanode_hash_info(ExplainState *es, int nbatch, int nbuckets_original, int nbatch_original,
int nbuckets, long spacePeakKb)
{
if (es->format != EXPLAIN_FORMAT_TEXT) {
ExplainPropertyLong("Hash Buckets", nbuckets, es);
ExplainPropertyLong("Original Hash Buckets", nbuckets_original, es);
ExplainPropertyLong("Hash Batches", nbatch, es);
ExplainPropertyLong("Original Hash Batches", nbatch_original, es);
ExplainPropertyLong("Peak Memory Usage", spacePeakKb, es);
} else if (is_detail == false && t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL &&
es->planinfo->m_staticInfo) {
if (nbatch_original != nbatch) {
appendStringInfo(es->planinfo->m_staticInfo->info_str,
" Buckets: %d (originally %d) Batches: %d (originally %d) Memory Usage: %ldkB\n",
nbuckets,
nbuckets_original,
nbatch,
nbatch_original,
spacePeakKb);
} else {
appendStringInfo(es->planinfo->m_staticInfo->info_str,
" Buckets: %d Batches: %d Memory Usage: %ldkB\n",
nbuckets,
nbatch,
spacePeakKb);
}
} else {
if (nbatch_original != nbatch) {
appendStringInfo(es->str,
" Buckets: %d (originally %d) Batches: %d (originally %d) Memory Usage: %ldkB\n",
nbuckets,
nbuckets_original,
nbatch,
nbatch_original,
spacePeakKb);
} else {
appendStringInfo(
es->str, " Buckets: %d Batches: %d Memory Usage: %ldkB\n", nbuckets, nbatch, spacePeakKb);
}
}
}
* Brief : Display LLVM info through explain performance syntax.
* Description : Display LLVM info through explain performance syntax.
* Input : fsstate, a ForeignScanState struct.
* es, a ExplainState struct.
* Output : none.
* Return Value : none.
* Notes : none.
*/
static void show_llvm_info(const PlanState* planstate, ExplainState* es)
{
if (!es->detail)
return;
AssertEreport(planstate != NULL && es != NULL, MOD_EXECUTOR, "unexpect null value");
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->format == EXPLAIN_FORMAT_TEXT) {
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr != NULL && instr->isLlvmOpt) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(i, 0, LLVM_OPTIMIZATION, true);
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_datanodeInfo->info_str, "%s ", node_name);
appendStringInfoString(es->planinfo->m_datanodeInfo->info_str, "(LLVM Optimized)\n");
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, " %s", node_name);
appendStringInfo(es->str, " (LLVM Optimized)\n");
}
}
}
} else {
bool first = true;
bool has_llvm = false;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL)
continue;
if (instr->isLlvmOpt) {
has_llvm = true;
if (first) {
ExplainOpenGroup("LLVM Detail", "LLVM Detail", false, es);
}
ExplainOpenGroup("Plan", NULL, true, es);
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
ExplainPropertyText("DN Name", node_name, es);
ExplainPropertyText("LLVM", "LLVM Optimized", es);
ExplainCloseGroup("Plan", NULL, true, es);
first = false;
}
}
if (has_llvm)
ExplainCloseGroup("LLVM Detail", "LLVM Detail", false, es);
}
}
}
* @Description: show datanode filenum and respill info
* @in es: current explainstate
* @in file_num: file num to show
* @in respill_time: respill times if any(>0) to show
* @in expand_times: hashtable expand times
* @return: void
*/
static void show_datanode_filenum_info(
ExplainState* es, int file_num, int respill_time, int expand_times, StringInfo info_str)
{
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (file_num > 0) {
appendStringInfo(info_str, "Temp File Num: %d", file_num);
if (respill_time > 0) {
appendStringInfo(info_str, ", Spill Time: %d", respill_time);
}
if (expand_times > 0)
appendStringInfo(info_str, ", Expand Times: %d", expand_times);
}
if (file_num == 0 && expand_times > 0) {
appendStringInfo(info_str, "Expand Times: %d", expand_times);
}
appendStringInfoChar(info_str, '\n');
} else {
ExplainPropertyLong("Temp File Num", file_num, es);
ExplainPropertyLong("Spill Time", respill_time, es);
ExplainPropertyLong("Expand Times", expand_times, es);
}
}
* @Description: show detail filenum
* @in es: current explainstate
* @in planstate: planstate info
* @in label_name: operator label name
* @return: void
*/
static void show_detail_filenum_info(const PlanState* planstate, ExplainState* es, const char* label_name)
{
Instrumentation* instr = NULL;
int max_file_num = 0;
int min_file_num = MIN_FILE_NUM;
bool is_execute = false;
bool is_writefile = false;
int datanode_size = 0;
int i = 0;
int j = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
int count_dn_writefile = 0;
if (u_sess->instr_cxt.global_instr)
datanode_size = u_sess->instr_cxt.global_instr->getInstruNodeNum();
if (es->detail) {
ExplainOpenGroup(label_name, label_name, false, es);
for (i = 0; i < datanode_size; i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
is_writefile = instr->sorthashinfo.hash_writefile;
int expand_times = instr->sorthashinfo.hashtable_expand_times;
if (is_writefile || expand_times > 0) {
ExplainOpenGroup("Plan", NULL, true, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL
&& es->planinfo->m_runtimeinfo != NULL && es->planinfo->m_staticInfo != NULL) {
es->planinfo->m_staticInfo->set_plan_name<false, true>();
es->planinfo->m_staticInfo->set_datanode_name(node_name, j, dop);
if (is_writefile) {
es->planinfo->m_runtimeinfo->put(i, j, HASH_FILENUM, instr->sorthashinfo.hash_FileNum);
}
show_datanode_filenum_info(es,
instr->sorthashinfo.hash_FileNum,
instr->sorthashinfo.hash_spillNum,
expand_times,
es->planinfo->m_staticInfo->info_str);
} else {
append_datanode_name(es, node_name, dop, j);
show_datanode_filenum_info(es,
instr->sorthashinfo.hash_FileNum,
instr->sorthashinfo.hash_spillNum,
expand_times,
es->str);
}
ExplainCloseGroup("Plan", NULL, true, es);
}
}
}
}
ExplainCloseGroup(label_name, label_name, false, es);
} else {
if (datanode_size > 0) {
int file_num = 0;
for (i = 0; i < datanode_size; i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
int count_dop_writefile = 0;
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
if (!is_execute) {
is_execute = true;
}
is_writefile = instr->sorthashinfo.hash_writefile;
if (is_writefile == true) {
if (count_dop_writefile++ == 0) {
count_dn_writefile++;
}
file_num = instr->sorthashinfo.hash_FileNum;
max_file_num = rtl::max(max_file_num, file_num);
min_file_num = rtl::min(min_file_num, file_num);
if (min_file_num == MIN_FILE_NUM) {
min_file_num = file_num;
}
}
}
}
}
if (count_dn_writefile < datanode_size) {
min_file_num = 0;
}
if (is_execute && max_file_num > 0) {
if (es->format != EXPLAIN_FORMAT_TEXT) {
ExplainPropertyLong("Max File Num", max_file_num, es);
ExplainPropertyLong("Min File Num", min_file_num, es);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Max File Num: %d Min File Num: %d\n", max_file_num, min_file_num);
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_PRETTY && es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Max File Num: %d Min File Num: %d\n",
max_file_num,
min_file_num);
}
}
}
}
}
}
* Show information on hashed_setop.
*/
static void show_setop_info(SetOpState* setopstate, ExplainState* es)
{
PlanState* planstate = (PlanState*)setopstate;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr && es->from_dn) {
show_detail_filenum_info(planstate, es, "Hash Setop Detail");
} else if (es->analyze) {
if (!es->from_dn) {
int filenum = 0;
int expand_times = 0;
if (IsA(setopstate, VecSetOpState)) {
setOpTbl* vechashSetopTbl = (setOpTbl*)((VecSetOpState*)setopstate)->vecSetOpInfo;
if (vechashSetopTbl != NULL)
filenum = vechashSetopTbl->getFileNum();
expand_times = setopstate->ps.instrument->sorthashinfo.hashtable_expand_times;
} else {
SetopWriteFileControl* TempFileControl = (SetopWriteFileControl*)setopstate->TempFileControl;
if (TempFileControl != NULL)
filenum = TempFileControl->filenum;
}
if (filenum > 0 || expand_times > 0) {
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL)
show_datanode_filenum_info(es, filenum, 0, expand_times, es->str);
else {
if (es->planinfo->m_runtimeinfo)
es->planinfo->m_runtimeinfo->put(-1, -1, HASH_FILENUM, filenum);
if (es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
show_datanode_filenum_info(es, filenum, 0, expand_times, es->planinfo->m_staticInfo->info_str);
}
}
}
}
}
}
* @Description: Show hashagg build and probe time info
* @in planstate: PlanState node.+ * @in es: Explain state.
*/
static void show_detail_execute_info(const PlanState* planstate, ExplainState* es)
{
Instrumentation* instr = NULL;
int datanode_size = 0;
int i = 0;
int j = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (u_sess->instr_cxt.global_instr)
datanode_size = u_sess->instr_cxt.global_instr->getInstruNodeNum();
if (!es->detail)
return;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_runtimeinfo) {
for (i = 0; i < datanode_size; i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
double build_time = instr->sorthashinfo.hashbuild_time;
double agg_time = instr->sorthashinfo.hashagg_time;
if (build_time > 0.0 || agg_time > 0.0) {
es->planinfo->m_staticInfo->set_plan_name<false, true>();
es->planinfo->m_staticInfo->set_datanode_name(node_name, j, dop);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Hashagg Build time: %.3f, Aggregation time: %.3f\n",
build_time * 1000.0,
agg_time * 1000.0);
}
}
}
}
}
}
* Show information on hash agg.
*/
static void show_hashAgg_info(AggState* hashaggstate, ExplainState* es)
{
PlanState* planstate = (PlanState*)hashaggstate;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr && es->from_dn) {
show_detail_filenum_info(planstate, es, "Hash Aggregate Detail");
show_detail_execute_info(planstate, es);
} else if (es->analyze) {
if (!es->from_dn) {
int filenum = 0;
int expand_times = 0;
double hashagg_build_time = 0.0;
double hashagg_agg_time = 0.0;
if (IsA(hashaggstate, VecAggState)) {
VecAgg* plan = (VecAgg*)(hashaggstate->ss.ps.plan);
if (plan->is_sonichash) {
SonicHashAgg* vechashAggTbl = (SonicHashAgg*)((VecAggState*)hashaggstate)->aggRun;
if (vechashAggTbl != NULL)
filenum = vechashAggTbl->getFileNum();
} else {
HashAggRunner* vechashAggTbl = (HashAggRunner*)((VecAggState*)hashaggstate)->aggRun;
if (vechashAggTbl != NULL)
filenum = vechashAggTbl->getFileNum();
}
expand_times = hashaggstate->ss.ps.instrument->sorthashinfo.hashtable_expand_times;
hashagg_build_time = hashaggstate->ss.ps.instrument->sorthashinfo.hashbuild_time;
hashagg_agg_time = hashaggstate->ss.ps.instrument->sorthashinfo.hashagg_time;
} else {
AggWriteFileControl* TempFileControl = (AggWriteFileControl*)hashaggstate->aggTempFileControl;
if (TempFileControl != NULL)
filenum = TempFileControl->filenum;
}
if (filenum > 0 || expand_times > 0) {
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL) {
show_datanode_filenum_info(es,
filenum,
planstate->instrument->sorthashinfo.hash_spillNum,
planstate->instrument->sorthashinfo.hashtable_expand_times,
es->str);
} else {
if (es->planinfo->m_runtimeinfo != NULL)
es->planinfo->m_runtimeinfo->put(-1, -1, HASH_FILENUM, filenum);
if (es->planinfo->m_staticInfo != NULL) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
show_datanode_filenum_info(es,
filenum,
planstate->instrument->sorthashinfo.hash_spillNum,
planstate->instrument->sorthashinfo.hashtable_expand_times,
es->planinfo->m_staticInfo->info_str);
}
}
}
if (hashagg_build_time > 0.0 || hashagg_agg_time > 0.0) {
if (es->planinfo && es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Hashagg Build time: %.3f, Aggregation time: %.3f\n",
hashagg_build_time * 1000.0,
hashagg_agg_time * 1000.0);
}
}
}
}
}
* Show information on hash buckets/batches.
*/
static void show_hash_info(HashState* hashstate, ExplainState* es)
{
HashJoinTable hashtable;
PlanState* planstate = NULL;
planstate = (PlanState*)hashstate;
int nbatch;
int nbatch_original;
int nbuckets;
int nbuckets_original;
long spacePeakKb = 0;
int max_nbatch = -1;
int min_nbatch = INT_MAX;
int max_nbatch_original = -1;
int min_nbatch_original = INT_MAX;
int max_nbuckets = -1;
int min_nbuckets = INT_MAX;
long max_spacePeakKb = -1;
long min_spacePeakKb = LONG_MAX;
bool is_execute = false;
AssertEreport(IsA(hashstate, HashState), MOD_EXECUTOR, "unexpect hashstate type");
hashtable = hashstate->hashtable;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->detail) {
ExplainOpenGroup("Hash Detail", "Hash Detail", false, es);
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
append_datanode_name(es, node_name, 1, 0);
spacePeakKb = (instr->sorthashinfo.spacePeak + 1023) / 1024;
nbatch = instr->sorthashinfo.nbatch;
nbatch_original = instr->sorthashinfo.nbatch_original;
nbuckets = instr->sorthashinfo.nbuckets;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo
&& es->planinfo->m_staticInfo && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(i, 0, HASH_BATCH, nbatch);
es->planinfo->m_runtimeinfo->put(i, 0, HASH_BATCH_ORIGNAL, nbatch_original);
es->planinfo->m_runtimeinfo->put(i, 0, HASH_BUCKET, nbuckets);
es->planinfo->m_runtimeinfo->put(i, 0, HASH_SPACE, spacePeakKb);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
show_datanode_hash_info<false>(es, nbatch, nbuckets, nbatch_original, nbuckets, spacePeakKb);
}
show_datanode_hash_info<true>(es, nbatch, nbuckets, nbatch_original, nbuckets, spacePeakKb);
ExplainCloseGroup("Plan", NULL, true, es);
}
ExplainCloseGroup("Hash Detail", "Hash Detail", false, es);
} else {
if (u_sess->instr_cxt.global_instr->getInstruNodeNum() > 0) {
SortHashInfo* psi = NULL;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr != NULL) {
if (!is_execute)
is_execute = true;
psi = &instr->sorthashinfo;
spacePeakKb = (instr->sorthashinfo.spacePeak + 1023) / 1024;
max_nbatch = rtl::max(max_nbatch, psi->nbatch);
min_nbatch = rtl::min(min_nbatch, psi->nbatch);
max_nbatch_original = rtl::max(max_nbatch_original, psi->nbatch_original);
min_nbatch_original = rtl::min(min_nbatch_original, psi->nbatch_original);
max_nbuckets = rtl::max(max_nbuckets, psi->nbuckets);
min_nbuckets = rtl::min(min_nbuckets, psi->nbuckets);
max_spacePeakKb = rtl::max(max_spacePeakKb, spacePeakKb);
min_spacePeakKb = rtl::min(min_spacePeakKb, spacePeakKb);
}
}
if (is_execute) {
AssertEreport(max_nbatch != -1 && min_nbatch != INT_MAX && max_nbatch_original != -1 &&
min_nbatch_original != INT_MAX && max_nbuckets != -1 && min_nbuckets != INT_MAX &&
max_spacePeakKb != -1 && min_spacePeakKb != LONG_MAX,
MOD_EXECUTOR,
"unexpect min/max value");
if (es->format != EXPLAIN_FORMAT_TEXT) {
ExplainPropertyLong("Max Hash Buckets", max_nbuckets, es);
ExplainPropertyLong("Min Hash Buckets", min_nbuckets, es);
ExplainPropertyLong("Max Hash Batches", max_nbatch, es);
ExplainPropertyLong("Min Hash Batches", min_nbatch, es);
ExplainPropertyLong("Max Original Hash Batches", max_nbatch_original, es);
ExplainPropertyLong("Min Original Hash Batches", min_nbatch_original, es);
ExplainPropertyLong("Max Peak Memory Usage", max_spacePeakKb, es);
ExplainPropertyLong("min Peak Memory Usage", min_spacePeakKb, es);
} else if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL) {
if (max_nbatch_original != max_nbatch) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str,
"Max Buckets: %d Max Batches: %d (max originally %d) Max Memory Usage: %ldkB\n",
max_nbuckets,
max_nbatch,
max_nbatch_original,
max_spacePeakKb);
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str,
"Min Buckets: %d Min Batches: %d (min originally %d) Min Memory Usage: %ldkB\n",
min_nbuckets,
min_nbatch,
min_nbatch_original,
min_spacePeakKb);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str,
"Max Buckets: %d Max Batches: %d Max Memory Usage: %ldkB\n",
max_nbuckets,
max_nbatch,
max_spacePeakKb);
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str,
"Min Buckets: %d Min Batches: %d Min Memory Usage: %ldkB\n",
min_nbuckets,
min_nbatch,
min_spacePeakKb);
}
} else if (es->planinfo->m_staticInfo) {
if (max_nbatch_original != max_nbatch) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Max Buckets: %d Max Batches: %d (max originally %d) Max Memory Usage: %ldkB\n",
max_nbuckets,
max_nbatch,
max_nbatch_original,
max_spacePeakKb);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Min Buckets: %d Min Batches: %d (min originally %d) Min Memory Usage: %ldkB\n",
min_nbuckets,
min_nbatch,
min_nbatch_original,
min_spacePeakKb);
} else {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Max Buckets: %d Max Batches: %d Max Memory Usage: %ldkB\n",
max_nbuckets,
max_nbatch,
max_spacePeakKb);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"Min Buckets: %d Min Batches: %d Min Memory Usage: %ldkB\n",
min_nbuckets,
min_nbatch,
min_spacePeakKb);
}
}
}
}
}
} else if (hashstate->ps.instrument) {
SortHashInfo hashinfo = hashstate->ps.instrument->sorthashinfo;
spacePeakKb = (hashinfo.spacePeak + 1023) / 1024;
nbatch = hashinfo.nbatch;
nbatch_original = hashinfo.nbatch_original;
nbuckets = hashinfo.nbuckets;
if (es->analyze) {
nbuckets_original = hashtable ? hashtable->nbuckets_original : nbuckets;
} else {
nbuckets_original = nbuckets;
}
if (es->wlm_statistics_plan_max_digit == NULL) {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoSpaces(es->str, es->indent * 2);
show_datanode_hash_info<false>(es, nbatch, nbuckets_original, nbatch_original, nbuckets, spacePeakKb);
}
}
}
inline static void show_buffers_info(
StringInfo infostr, bool has_shared, bool has_local, bool has_temp, const BufferUsage* usage)
{
appendStringInfoString(infostr, "(Buffers:");
if (has_shared) {
appendStringInfoString(infostr, " shared");
if (usage->shared_blks_hit > 0)
appendStringInfo(infostr, " hit=%ld", usage->shared_blks_hit);
if (usage->shared_blks_read > 0)
appendStringInfo(infostr, " read=%ld", usage->shared_blks_read);
if (usage->shared_blks_dirtied > 0)
appendStringInfo(infostr, " dirtied=%ld", usage->shared_blks_dirtied);
if (usage->shared_blks_written > 0)
appendStringInfo(infostr, " written=%ld", usage->shared_blks_written);
}
if (has_local) {
appendStringInfoString(infostr, " local");
if (usage->local_blks_hit > 0)
appendStringInfo(infostr, " hit=%ld", usage->local_blks_hit);
if (usage->local_blks_read > 0)
appendStringInfo(infostr, " read=%ld", usage->local_blks_read);
if (usage->local_blks_dirtied > 0)
appendStringInfo(infostr, " dirtied=%ld", usage->local_blks_dirtied);
if (usage->local_blks_written > 0)
appendStringInfo(infostr, " written=%ld", usage->local_blks_written);
if (has_temp)
appendStringInfoChar(infostr, ',');
}
if (has_temp) {
appendStringInfoString(infostr, " temp");
if (usage->temp_blks_read > 0)
appendStringInfo(infostr, " read=%ld", usage->temp_blks_read);
if (usage->temp_blks_written > 0)
appendStringInfo(infostr, " written=%ld", usage->temp_blks_written);
}
appendStringInfoString(infostr, ")\n");
}
static void put_buffers(ExplainState* es, const BufferUsage* usage, int nodeIdx, int smpIdx)
{
bool has_timing = false;
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, SHARED_BLK_HIT, Int64GetDatum(usage->shared_blks_hit));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, SHARED_BLK_READ, Int64GetDatum(usage->shared_blks_read));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, SHARED_BLK_DIRTIED, Int64GetDatum(usage->shared_blks_dirtied));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, SHARED_BLK_WRITTEN, Int64GetDatum(usage->shared_blks_written));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, LOCAL_BLK_HIT, Int64GetDatum(usage->local_blks_hit));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, LOCAL_BLK_READ, Int64GetDatum(usage->local_blks_read));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, LOCAL_BLK_DIRTIED, Int64GetDatum(usage->local_blks_dirtied));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, LOCAL_BLK_WRITTEN, Int64GetDatum(usage->local_blks_written));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, TEMP_BLK_READ, Int64GetDatum(usage->temp_blks_read));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, TEMP_BLK_WRITTEN, Int64GetDatum(usage->temp_blks_written));
has_timing = (!INSTR_TIME_IS_ZERO(usage->blk_read_time) || !INSTR_TIME_IS_ZERO(usage->blk_write_time));
if (has_timing) {
double blk_time;
if (!INSTR_TIME_IS_ZERO(usage->blk_read_time)) {
blk_time = INSTR_TIME_GET_MILLISEC(usage->blk_read_time);
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, BLK_READ_TIME, Float8GetDatum(blk_time));
}
if (!INSTR_TIME_IS_ZERO(usage->blk_write_time)) {
blk_time = INSTR_TIME_GET_MILLISEC(usage->blk_write_time);
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, BLK_WRITE_TIME, Float8GetDatum(blk_time));
}
}
}
static void show_datanode_vecjoin_info(
ExplainState* es, StringInfo instr, int64 file_number, int spillNum, long spillSize, long space, bool is_writefile)
{
if (is_writefile == false) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(instr, "Memory Used : %ldkB\n", space);
} else {
ExplainPropertyLong("Memory Used", space, es);
}
} else {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (spillNum > 0)
appendStringInfo(instr,
"Temp File Num: %ld, Spill Time: %d, Total Written Disk IO: %ldKB\n",
file_number,
spillNum,
spillSize);
else
appendStringInfo(instr, "Temp File Num: %ld, Total Written Disk IO: %ldKB\n", file_number, spillSize);
} else {
ExplainPropertyLong("Temp File Num", file_number, es);
ExplainPropertyLong("Spill Time", spillNum, es);
ExplainPropertyLong("Total Written Disk IO", spillSize, es);
}
}
}
static void show_datanode_sonicjoin_info(
ExplainState* es, StringInfo instr, int partNum, int spillNum, long space, bool is_writefile)
{
if (is_writefile == false) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(instr, "Memory Used : %ldkB\n", space);
} else {
ExplainPropertyLong("Memory Used", space, es);
}
} else {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (spillNum > 0)
appendStringInfo(instr, "Partition Num: %d, Spill Times: %d\n", partNum, spillNum);
else
appendStringInfo(instr, "Partition Num: %d\n", partNum);
} else {
ExplainPropertyLong("Partition Num", partNum, es);
ExplainPropertyLong("Spill Times", spillNum, es);
}
}
}
static void show_datanode_sonicjoin_detail_info(ExplainState* es, StringInfo instr, bool buildside, int partNum,
int fileNum, long spillSize, long spillSizeMin, long spillSizeMax)
{
const char* side = buildside ? "Inner Partition" : "Outer Partition";
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(instr,
"%s Spill Num: %d, "
"Temp File Num: %d, "
"Written Disk IO: %ldKB [%ldKB %ldKB]\n",
side,
partNum,
fileNum,
spillSize,
spillSizeMin,
spillSizeMax);
} else {
ExplainOpenGroup(side, side, false, es);
ExplainPropertyLong("Spill Num", partNum, es);
ExplainPropertyLong("Temp File Num", fileNum, es);
ExplainPropertyLong("Total Written Disk IO", spillSize, es);
ExplainPropertyLong("Min Partition Written Disk IO", spillSizeMin, es);
ExplainPropertyLong("Max Partition Written Disk IO", spillSizeMax, es);
ExplainCloseGroup(side, side, false, es);
}
}
static void show_vechash_info(VecHashJoinState* hashstate, ExplainState* es)
{
PlanState* planstate = (PlanState*)hashstate;
int64 max_file_num = 0;
int min_file_num = MIN_FILE_NUM;
bool is_execute = false;
long spaceUsed, spillSize;
int64 file_num;
long max_spaceused = 0;
long min_spaceused = u_sess->attr.attr_memory.work_mem;
bool is_writefile = false;
int spillNum;
int partNum, innerPartNum, outerPartNum;
int innerFileNum, outerFileNum;
long spillInnerSize, spillInnerSizeMax, spillInnerSizeMin;
long spillOuterSize, spillOuterSizeMax, spillOuterSizeMin;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->detail) {
ExplainOpenGroup("VecHashJoin Detail", "VecHashJoin Detail", false, es);
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL)
continue;
is_writefile = instr->sorthashinfo.hash_writefile;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
spaceUsed = (instr->sorthashinfo.spaceUsed + 1023) / 1024;
spillSize = (instr->sorthashinfo.spill_size + 1023) / 1024;
file_num = instr->sorthashinfo.hash_FileNum;
spillNum = instr->sorthashinfo.hash_spillNum;
partNum = instr->sorthashinfo.hash_partNum;
innerPartNum = instr->sorthashinfo.spill_innerPartNum;
outerPartNum = instr->sorthashinfo.spill_outerPartNum;
innerFileNum = instr->sorthashinfo.hash_innerFileNum;
outerFileNum = instr->sorthashinfo.hash_outerFileNum;
spillInnerSize = (instr->sorthashinfo.spill_innerSize + 1023) / 1024;
spillInnerSizeMax = (instr->sorthashinfo.spill_innerSizePartMax + 1023) / 1024;
spillInnerSizeMin = (instr->sorthashinfo.spill_innerSizePartMin + 1023) / 1024;
spillOuterSize = (instr->sorthashinfo.spill_outerSize + 1023) / 1024;
spillOuterSizeMax = (instr->sorthashinfo.spill_outerSizePartMax + 1023) / 1024;
spillOuterSizeMin = (instr->sorthashinfo.spill_outerSizePartMin + 1023) / 1024;
double build_time = instr->sorthashinfo.hashbuild_time;
double probe_time = instr->sorthashinfo.hashagg_time;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL &&
es->planinfo != NULL && es->planinfo->m_staticInfo != NULL) {
if (is_writefile)
es->planinfo->m_runtimeinfo->put(i, 0, HASH_FILENUM, file_num);
else
es->planinfo->m_runtimeinfo->put(i, 0, HASH_SPACE, spaceUsed);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
if (((HashJoin*)planstate->plan)->isSonicHash) {
show_datanode_sonicjoin_info(
es, es->planinfo->m_staticInfo->info_str, partNum, spillNum, spaceUsed, is_writefile);
if (is_writefile) {
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
show_datanode_sonicjoin_detail_info(es,
es->planinfo->m_staticInfo->info_str,
true,
innerPartNum,
innerFileNum,
spillInnerSize,
spillInnerSizeMin,
spillInnerSizeMax);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s ", node_name);
show_datanode_sonicjoin_detail_info(es,
es->planinfo->m_staticInfo->info_str,
false,
outerPartNum,
outerFileNum,
spillOuterSize,
spillOuterSizeMin,
spillOuterSizeMax);
}
} else
show_datanode_vecjoin_info(es,
es->planinfo->m_staticInfo->info_str,
file_num,
spillNum,
spillSize,
spaceUsed,
is_writefile);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str, "%s", node_name);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
" Hashjoin Build time: %.3f, Probe time: %.3f\n",
build_time * 1000.0,
probe_time * 1000.0);
} else {
append_datanode_name(es, node_name, 1, 0);
if (((HashJoin*)planstate->plan)->isSonicHash) {
show_datanode_sonicjoin_info(es, es->str, partNum, spillNum, spaceUsed, is_writefile);
if (is_writefile) {
append_datanode_name(es, node_name, 1, 0);
show_datanode_sonicjoin_detail_info(es,
es->str,
true,
innerPartNum,
innerFileNum,
spillInnerSize,
spillInnerSizeMin,
spillInnerSizeMax);
append_datanode_name(es, node_name, 1, 0);
show_datanode_sonicjoin_detail_info(es,
es->str,
false,
outerPartNum,
outerFileNum,
spillOuterSize,
spillOuterSizeMin,
spillOuterSizeMax);
}
} else {
show_datanode_vecjoin_info(es, es->str, file_num, spillNum, spillSize, spaceUsed, is_writefile);
}
}
ExplainCloseGroup("Plan", NULL, true, es);
}
ExplainCloseGroup("VecHashJoin Detail", "VecHashJoin Detail", false, es);
} else {
if (u_sess->instr_cxt.global_instr->getInstruNodeNum() > 0) {
file_num = 0;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr != NULL) {
if (!is_execute)
is_execute = true;
file_num = instr->sorthashinfo.hash_FileNum;
spaceUsed = (instr->sorthashinfo.spaceUsed + 1023) / 1024;
max_file_num = rtl::max(max_file_num, file_num);
min_file_num = rtl::min((int64)min_file_num, file_num);
if (min_file_num == MIN_FILE_NUM)
min_file_num = file_num;
max_spaceused = rtl::max(max_spaceused, spaceUsed);
min_spaceused = rtl::min(min_spaceused, spaceUsed);
}
}
if (is_execute) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL) {
* Here, we adjust output information and here our output will be used
* major showing whether writing to disk occurs or not.
* 1, min_file_num and max_file_num are both equal to 0.
* It means All DNs' work mem can hold themselves hash table in memory,
* then we only print out memory usage related information.
* 2, min_spaceused and max_spaceused are both equal to 0.
* It means all the DNs have written out the hash table content to disk,
* then we only print out min and max file num information.
* 3, If file num and spaceused are not both equal to 0.
* It means some DNs' work mem are not enough to hold themselves hash table in memory,
* here we print out only file num related informaiton to show the writing out operation
* occurs. 4, All above four variables equal to 0, this situation could occur when the hash
* table is empty.
*/
if (min_file_num != 0 || max_file_num != 0) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Max File Num: %ld\n", max_file_num);
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Min File Num: %d\n", min_file_num);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Max Memory Used : %ldkB\n", max_spaceused);
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Min Memory Used : %ldkB\n", min_spaceused);
}
} else if (es->planinfo->m_staticInfo) {
if (min_file_num != 0 || max_file_num != 0) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Max File Num: %ld\n", max_file_num);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Min File Num: %d\n", min_file_num);
} else {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Max Memory Used : %ldkB\n", max_spaceused);
es->planinfo->m_staticInfo->set_plan_name<false, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Min Memory Used : %ldkB\n", min_spaceused);
}
}
} else {
ExplainPropertyLong("Min Memory Used", min_spaceused, es);
ExplainPropertyLong("Max Memory Used", max_spaceused, es);
ExplainPropertyLong("Min File Num", min_file_num, es);
ExplainPropertyLong("Max File Num", max_file_num, es);
}
}
}
}
}
}
static void show_startwith_dfx(StartWithOpState* swstate, ExplainState* es)
{
List* result = NIL;
IterationStats* iters = &(swstate->iterStats);
if (!es->verbose || iters->currentStartTime.tv_sec == 0) {
return;
}
const char *qlabel = "Start With Iteration Statistics";
bool rotated = (iters->totalIters > SW_LOG_ROWS_FULL) ? true : false;
int offset = rotated ? iters->totalIters % SW_LOG_ROWS_HALF : 0;
StringInfo si = makeStringInfo();
for (int i = 0; i < SW_LOG_ROWS_FULL && i < iters->totalIters; i++) {
int ri = (i >= SW_LOG_ROWS_HALF && rotated) ?
SW_LOG_ROWS_HALF + (i + offset) % SW_LOG_ROWS_HALF : i;
double total_time = (iters->endTimeBuf[ri].tv_sec - iters->startTimeBuf[ri].tv_sec) / 0.001 +
(iters->endTimeBuf[ri].tv_usec - iters->startTimeBuf[ri].tv_usec) * 0.001;
appendStringInfo(
si, "\nIteration: %d, Row(s): %ld, Time Cost: %.3lf ms",
iters->levelBuf[ri], iters->rowCountBuf[ri], total_time);
result = lappend(result, pstrdup(si->data));
resetStringInfo(si);
}
DestroyStringInfo(si);
ExplainPropertyListPostPlanTree(qlabel, result, es);
list_free_deep(result);
}
static void show_recursive_info(RecursiveUnionState* rustate, ExplainState* es)
{
PlanState* planstate = (PlanState*)rustate;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->detail && t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL) {
ExplainOpenGroup("RecursiveUnion Detail", "RecursiveUnion Detail", false, es);
int niters = 0;
bool has_reach_limit = false;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL) {
continue;
}
niters = Max(niters, instr->recursiveInfo.niters);
if (instr->recursiveInfo.has_reach_limit)
has_reach_limit = true;
}
if (niters == 0) {
ExplainCloseGroup("RecursiveUnion Detail", "RecursiveUnion Detail", false, es);
return;
}
appendStringInfo(
es->planinfo->m_recursiveInfo->info_str, "%3d --Recursive Union\n", planstate->plan->plan_node_id);
appendStringInfoSpaces(es->planinfo->m_recursiveInfo->info_str, 8);
appendStringInfo(es->planinfo->m_recursiveInfo->info_str, "Iteration times: %d\n", niters - 1);
appendStringInfoSpaces(es->planinfo->m_recursiveInfo->info_str, 8);
appendStringInfo(es->planinfo->m_recursiveInfo->info_str, "Iteration[0] (Step 0 None-Recursive)\n");
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL) {
continue;
}
appendStringInfoSpaces(es->planinfo->m_recursiveInfo->info_str, 16);
appendStringInfo(es->planinfo->m_recursiveInfo->info_str,
"%s return tuples: %lu\n",
node_name,
instr->recursiveInfo.iter_ntuples[0]);
}
for (int iteridx = 1; iteridx < niters; iteridx++) {
appendStringInfoSpaces(es->planinfo->m_recursiveInfo->info_str, 8);
appendStringInfo(es->planinfo->m_recursiveInfo->info_str,
"Iteration[%d] (Step %d Recursive[%d])%s\n",
iteridx,
iteridx,
iteridx,
(iteridx == niters - 1 && !has_reach_limit) ? " --- finish" : "");
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL) {
continue;
}
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
appendStringInfoSpaces(es->planinfo->m_recursiveInfo->info_str, 16);
appendStringInfo(es->planinfo->m_recursiveInfo->info_str,
"%s return tuples: %lu\n",
node_name,
instr->recursiveInfo.iter_ntuples[iteridx]);
}
}
ExplainCloseGroup("RecursiveUnion Detail", "RecursiveUnion Detail", false, es);
}
}
}
static void show_datanode_buffers(ExplainState* es, PlanState* planstate)
{
Instrumentation* instr = NULL;
int nodeNum = u_sess->instr_cxt.global_instr->getInstruNodeNum();
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
int i = 0;
int j = 0;
if (es->detail) {
ExplainOpenGroup("Buffers In Detail", "Buffers In Detail", false, es);
for (i = 0; i < nodeNum; i++) {
for (j = 0; j < dop; j++) {
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr == NULL)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
append_datanode_name(es, node_name, dop, j);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
es->planinfo->m_datanodeInfo->set_datanode_name(node_name, j, dop);
}
show_buffers(es, es->str, instr, true, i, j, node_name);
ExplainCloseGroup("Plan", NULL, true, es);
}
}
ExplainCloseGroup("Buffers In Detail", "Buffers In Detail", false, es);
} else {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_IOInfo)
show_analyze_buffers(es, planstate, es->planinfo->m_IOInfo->info_str, nodeNum);
else
show_analyze_buffers(es, planstate, es->str, nodeNum);
}
}
static void show_analyze_buffers(ExplainState* es, const PlanState* planstate, StringInfo infostr, int nodeNum)
{
bool has_shared = false;
bool has_local = false;
bool has_temp = false;
bool has_timing = false;
long shared_blks_hit_max = -1;
long shared_blks_hit_min = LONG_MAX;
long shared_blks_read_max = -1;
long shared_blks_read_min = LONG_MAX;
long shared_blks_dirtied_max = -1;
long shared_blks_dirtied_min = LONG_MAX;
long shared_blks_written_max = -1;
long shared_blks_written_min = LONG_MAX;
long local_blks_hit_max = -1;
long local_blks_hit_min = LONG_MAX;
long local_blks_read_max = -1;
long local_blks_read_min = LONG_MAX;
long local_blks_dirtied_max = -1;
long local_blks_dirtied_min = LONG_MAX;
long local_blks_written_max = -1;
long local_blks_written_min = LONG_MAX;
long temp_blks_read_max = -1;
long temp_blks_read_min = LONG_MAX;
long temp_blks_written_max = -1;
long temp_blks_written_min = LONG_MAX;
instr_time blk_read_time_max, blk_read_time_min;
instr_time blk_write_time_max, blk_write_time_min;
bool is_execute = false;
Instrumentation* instr = NULL;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
INSTR_TIME_SET_ZERO(blk_read_time_max);
INSTR_TIME_SET_ZERO(blk_write_time_max);
INSTR_TIME_INITIAL_MIN(blk_read_time_min);
INSTR_TIME_INITIAL_MIN(blk_write_time_min);
if (nodeNum > 0) {
for (int i = 0; i < nodeNum; i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
for (int j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
BufferUsage* pbu = &instr->bufusage;
if (!is_execute) {
is_execute = true;
}
if (!has_shared)
has_shared = (pbu->shared_blks_hit > 0 || pbu->shared_blks_read > 0 ||
pbu->shared_blks_dirtied > 0 || pbu->shared_blks_written > 0);
shared_blks_read_max = rtl::max(shared_blks_read_max, pbu->shared_blks_read);
shared_blks_read_min = rtl::min(shared_blks_read_min, pbu->shared_blks_read);
shared_blks_hit_max = rtl::max(shared_blks_hit_max, pbu->shared_blks_hit);
shared_blks_hit_min = rtl::min(shared_blks_hit_min, pbu->shared_blks_hit);
shared_blks_dirtied_max = rtl::max(shared_blks_dirtied_max, pbu->shared_blks_dirtied);
shared_blks_dirtied_min = rtl::min(shared_blks_dirtied_min, pbu->shared_blks_dirtied);
shared_blks_written_max = rtl::max(shared_blks_written_max, pbu->shared_blks_written);
shared_blks_written_min = rtl::min(shared_blks_written_min, pbu->shared_blks_written);
if (!has_local)
has_local = (pbu->local_blks_hit > 0 || pbu->local_blks_read > 0 ||
pbu->local_blks_dirtied > 0 || pbu->local_blks_written > 0);
local_blks_read_max = rtl::max(local_blks_read_max, pbu->local_blks_read);
local_blks_read_min = rtl::min(local_blks_read_min, pbu->local_blks_read);
local_blks_hit_max = rtl::max(local_blks_hit_max, pbu->local_blks_hit);
local_blks_hit_min = rtl::min(local_blks_hit_min, pbu->local_blks_hit);
local_blks_dirtied_max = rtl::max(local_blks_dirtied_max, pbu->local_blks_dirtied);
local_blks_dirtied_min = rtl::min(local_blks_dirtied_min, pbu->local_blks_dirtied);
local_blks_written_max = rtl::max(local_blks_written_max, pbu->local_blks_written);
local_blks_written_min = rtl::min(local_blks_written_min, pbu->local_blks_written);
if (!has_temp)
has_temp = (pbu->temp_blks_read > 0 || pbu->temp_blks_written > 0);
temp_blks_read_max = rtl::max(temp_blks_read_max, pbu->temp_blks_read);
temp_blks_read_min = rtl::min(temp_blks_read_min, pbu->temp_blks_read);
temp_blks_written_max = rtl::max(temp_blks_written_max, pbu->temp_blks_written);
temp_blks_written_min = rtl::min(temp_blks_written_min, pbu->temp_blks_written);
if (!has_timing)
has_timing =
(!INSTR_TIME_IS_ZERO(pbu->blk_read_time) || !INSTR_TIME_IS_ZERO(pbu->blk_write_time));
if (!INSTR_TIME_IS_BIGGER(blk_read_time_max, pbu->blk_read_time))
blk_read_time_max = pbu->blk_read_time;
if (INSTR_TIME_IS_BIGGER(blk_read_time_min, pbu->blk_read_time))
blk_read_time_min = pbu->blk_read_time;
if (!INSTR_TIME_IS_BIGGER(blk_write_time_max, pbu->blk_write_time))
blk_write_time_max = pbu->blk_write_time;
if (INSTR_TIME_IS_BIGGER(blk_write_time_min, pbu->blk_write_time))
blk_write_time_min = pbu->blk_write_time;
}
}
}
if (is_execute) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (has_shared || has_local || has_temp) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_IOInfo) {
es->planinfo->m_IOInfo->set_plan_name<true, true>();
appendStringInfoString(es->planinfo->m_IOInfo->info_str, "Buffers:");
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, "Buffers:");
}
if (has_shared) {
appendStringInfoString(infostr, " shared");
if (shared_blks_hit_max > 0)
appendStringInfo(infostr, " max hit=%ld", shared_blks_hit_max);
if (shared_blks_hit_min > 0 && shared_blks_hit_min < LONG_MAX)
appendStringInfo(infostr, " min hit=%ld", shared_blks_hit_min);
if (shared_blks_read_max > 0)
appendStringInfo(infostr, " max read=%ld", shared_blks_read_max);
if (shared_blks_read_min > 0 && shared_blks_read_min < LONG_MAX)
appendStringInfo(infostr, " min read=%ld", shared_blks_read_min);
if (shared_blks_dirtied_max > 0)
appendStringInfo(infostr, " max dirtied=%ld", shared_blks_dirtied_max);
if (shared_blks_dirtied_min > 0 && shared_blks_dirtied_min < LONG_MAX)
appendStringInfo(infostr, " min dirtied=%ld", shared_blks_dirtied_min);
if (shared_blks_written_max > 0)
appendStringInfo(infostr, " max written=%ld", shared_blks_written_max);
if (shared_blks_written_min > 0 && shared_blks_written_min < LONG_MAX)
appendStringInfo(infostr, " min written=%ld", shared_blks_written_min);
if (has_local || has_temp)
appendStringInfoChar(infostr, ',');
}
if (has_local) {
appendStringInfoString(infostr, " local");
if (local_blks_hit_max > 0)
appendStringInfo(infostr, " max hit=%ld", local_blks_hit_max);
if (local_blks_hit_min > 0 && local_blks_hit_min < LONG_MAX)
appendStringInfo(infostr, " min hit=%ld", local_blks_hit_min);
if (local_blks_read_max > 0)
appendStringInfo(infostr, " max read=%ld", local_blks_read_max);
if (local_blks_read_min > 0 && local_blks_read_min < LONG_MAX)
appendStringInfo(infostr, " min read=%ld", local_blks_read_min);
if (local_blks_dirtied_max > 0)
appendStringInfo(infostr, " max dirtied=%ld", local_blks_dirtied_max);
if (local_blks_dirtied_min > 0 && local_blks_dirtied_min < LONG_MAX)
appendStringInfo(infostr, " min dirtied=%ld", local_blks_dirtied_min);
if (local_blks_written_max > 0)
appendStringInfo(infostr, " max written=%ld", local_blks_written_max);
if (local_blks_written_min > 0 && local_blks_written_min < LONG_MAX)
appendStringInfo(infostr, " min written=%ld", local_blks_written_min);
if (has_temp)
appendStringInfoChar(infostr, ',');
}
if (has_temp) {
appendStringInfoString(infostr, " temp");
if (temp_blks_read_max > 0)
appendStringInfo(infostr, " max read=%ld", temp_blks_read_max);
if (temp_blks_read_min > 0 && temp_blks_read_min < LONG_MAX)
appendStringInfo(infostr, " min read=%ld", temp_blks_read_min);
if (temp_blks_written_max > 0)
appendStringInfo(infostr, " max written=%ld", temp_blks_written_max);
if (temp_blks_written_min > 0 && temp_blks_written_min < LONG_MAX)
appendStringInfo(infostr, " min written=%ld", temp_blks_written_min);
}
appendStringInfoChar(infostr, '\n');
}
if (has_timing) {
appendStringInfoSpaces(infostr, es->indent * 2);
appendStringInfoString(infostr, "I/O Timings:");
if (!INSTR_TIME_IS_ZERO(blk_read_time_max))
appendStringInfo(infostr, " max read=%.3f", INSTR_TIME_GET_MILLISEC(blk_read_time_max));
if (!INSTR_TIME_IS_ZERO(blk_read_time_min) && !INSTR_TIME_IS_INTMAX(blk_read_time_min))
appendStringInfo(infostr, " min read=%.3f", INSTR_TIME_GET_MILLISEC(blk_read_time_min));
if (!INSTR_TIME_IS_ZERO(blk_write_time_max))
appendStringInfo(infostr, " max write=%.3f", INSTR_TIME_GET_MILLISEC(blk_write_time_max));
if (!INSTR_TIME_IS_ZERO(blk_write_time_min) && !INSTR_TIME_IS_INTMAX(blk_write_time_min))
appendStringInfo(infostr, " min write=%.3f", INSTR_TIME_GET_MILLISEC(blk_write_time_min));
appendStringInfoChar(infostr, '\n');
}
} else {
ExplainPropertyLong("Max Shared Hit Blocks", shared_blks_hit_max, es);
ExplainPropertyLong("Min Shared Hit Blocks", shared_blks_hit_min, es);
ExplainPropertyLong("Max Shared Read Blocks", shared_blks_read_max, es);
ExplainPropertyLong("Min Shared Read Blocks", shared_blks_read_min, es);
ExplainPropertyLong("Max Shared Dirtied Blocks", shared_blks_dirtied_max, es);
ExplainPropertyLong("Min Shared Dirtied Blocks", shared_blks_dirtied_min, es);
ExplainPropertyLong("Max Shared Written Blocks", shared_blks_written_max, es);
ExplainPropertyLong("Max Shared Written Blocks", shared_blks_written_min, es);
ExplainPropertyLong("Max Local Hit Blocks", local_blks_hit_max, es);
ExplainPropertyLong("Min Local Hit Blocks", local_blks_hit_min, es);
ExplainPropertyLong("Max Local Read Blocks", local_blks_read_max, es);
ExplainPropertyLong("Min Local Read Blocks", local_blks_read_min, es);
ExplainPropertyLong("Max Local Dirtied Blocks", local_blks_dirtied_max, es);
ExplainPropertyLong("Min Local Dirtied Blocks", local_blks_dirtied_min, es);
ExplainPropertyLong("Max Local Written Blocks", local_blks_written_max, es);
ExplainPropertyLong("Min Local Written Blocks", local_blks_written_min, es);
ExplainPropertyLong("Max Temp Read Blocks", temp_blks_read_max, es);
ExplainPropertyLong("Min Temp Read Blocks", temp_blks_read_min, es);
ExplainPropertyLong("Max Temp Written Blocks", temp_blks_written_max, es);
ExplainPropertyLong("Min Temp Written Blocks", temp_blks_written_min, es);
ExplainPropertyFloat("Max IO Read Time", INSTR_TIME_GET_MILLISEC(blk_read_time_max), 3, es);
ExplainPropertyFloat("Min IO Read Time", INSTR_TIME_GET_MILLISEC(blk_read_time_min), 3, es);
ExplainPropertyFloat("Max IO Write Time", INSTR_TIME_GET_MILLISEC(blk_write_time_max), 3, es);
ExplainPropertyFloat("Min IO Write Time", INSTR_TIME_GET_MILLISEC(blk_write_time_min), 3, es);
}
}
}
}
static void show_buffers(ExplainState* es, StringInfo infostr, const Instrumentation* instrument, bool is_datanode,
int nodeIdx, int smpIdx, const char* node_name)
{
const BufferUsage* usage = &instrument->bufusage;
if (es->format == EXPLAIN_FORMAT_TEXT) {
bool has_shared = (usage->shared_blks_hit > 0 || usage->shared_blks_read > 0 ||
usage->shared_blks_dirtied > 0 || usage->shared_blks_written > 0);
bool has_local = (usage->local_blks_hit > 0 || usage->local_blks_read > 0 || usage->local_blks_dirtied > 0 ||
usage->local_blks_written > 0);
bool has_temp = (usage->temp_blks_read > 0 || usage->temp_blks_written > 0);
bool has_timing = (!INSTR_TIME_IS_ZERO(usage->blk_read_time) || !INSTR_TIME_IS_ZERO(usage->blk_write_time));
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->detail) {
put_buffers(es, usage, nodeIdx, smpIdx);
if (node_name == NULL)
appendStringInfoSpaces(es->planinfo->m_datanodeInfo->info_str, 8);
if (has_shared || has_temp) {
show_buffers_info(es->planinfo->m_datanodeInfo->info_str, has_shared, has_local, has_temp, usage);
} else {
appendStringInfo(es->planinfo->m_datanodeInfo->info_str, "(Buffers: 0)\n");
}
return;
}
if (has_shared || has_local || has_temp) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_IOInfo) {
es->planinfo->m_IOInfo->set_plan_name<true, true>();
appendStringInfoString(es->planinfo->m_IOInfo->info_str, "Buffers:");
}
if (!is_datanode)
appendStringInfoSpaces(es->str, es->indent * 2);
show_buffers_info(infostr, has_shared, has_local, has_temp, usage);
} else if (is_datanode) {
appendStringInfo(infostr, get_execute_mode(es, nodeIdx) ? "(Buffers: 0)\n" : "(Buffers: unknown)\n");
}
if (has_timing) {
appendStringInfoSpaces(infostr, es->indent * 2);
appendStringInfoString(infostr, "I/O Timings:");
if (!INSTR_TIME_IS_ZERO(usage->blk_read_time))
appendStringInfo(infostr, " read=%.3f", INSTR_TIME_GET_MILLISEC(usage->blk_read_time));
if (!INSTR_TIME_IS_ZERO(usage->blk_write_time))
appendStringInfo(infostr, " write=%.3f", INSTR_TIME_GET_MILLISEC(usage->blk_write_time));
appendStringInfoChar(infostr, '\n');
}
} else {
ExplainPropertyLong("Shared Hit Blocks", usage->shared_blks_hit, es);
ExplainPropertyLong("Shared Read Blocks", usage->shared_blks_read, es);
ExplainPropertyLong("Shared Dirtied Blocks", usage->shared_blks_dirtied, es);
ExplainPropertyLong("Shared Written Blocks", usage->shared_blks_written, es);
ExplainPropertyLong("Local Hit Blocks", usage->local_blks_hit, es);
ExplainPropertyLong("Local Read Blocks", usage->local_blks_read, es);
ExplainPropertyLong("Local Dirtied Blocks", usage->local_blks_dirtied, es);
ExplainPropertyLong("Local Written Blocks", usage->local_blks_written, es);
ExplainPropertyLong("Temp Read Blocks", usage->temp_blks_read, es);
ExplainPropertyLong("Temp Written Blocks", usage->temp_blks_written, es);
ExplainPropertyFloat("IO Read Time", INSTR_TIME_GET_MILLISEC(usage->blk_read_time), 3, es);
ExplainPropertyFloat("IO Write Time", INSTR_TIME_GET_MILLISEC(usage->blk_write_time), 3, es);
}
}
* Calculate the child plan's exclusive cpu cycles/inclusive cpu cycles/
* left child processed rows/right child processedrows.
*/
template <bool datanode>
static void show_child_cpu_cycles_and_rows(PlanState* planstate, int idx, int smpIdx, double* outerCycles,
double* innerCycles, uint64* outterRows, uint64* innerRows)
{
PlanState* outerChild = NULL;
PlanState* innerChild = NULL;
Instrumentation* tmp_instr = NULL;
Plan* plan = planstate->plan;
int outer_smp_idx = 0;
int inner_smp_idx = 0;
switch (nodeTag(plan)) {
case T_ModifyTable:
case T_VecModifyTable:
CalCPUMemberNode<datanode>(((ModifyTable*)plan)->plans,
((ModifyTableState*)planstate)->mt_plans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_VecAppend:
case T_Append:
CalCPUMemberNode<datanode>(((Append*)plan)->appendplans,
((AppendState*)planstate)->appendplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_MergeAppend:
CalCPUMemberNode<datanode>(((MergeAppend*)plan)->mergeplans,
((MergeAppendState*)planstate)->mergeplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_BitmapAnd:
CalCPUMemberNode<datanode>(((BitmapAnd*)plan)->bitmapplans,
((BitmapAndState*)planstate)->bitmapplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_BitmapOr:
CalCPUMemberNode<datanode>(((BitmapOr*)plan)->bitmapplans,
((BitmapOrState*)planstate)->bitmapplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_CStoreIndexAnd:
CalCPUMemberNode<datanode>(((CStoreIndexAnd*)plan)->bitmapplans,
((BitmapAndState*)planstate)->bitmapplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_CStoreIndexOr:
CalCPUMemberNode<datanode>(((CStoreIndexOr*)plan)->bitmapplans,
((BitmapOrState*)planstate)->bitmapplans,
idx,
smpIdx,
outerCycles,
outterRows);
break;
case T_SubqueryScan:
case T_VecSubqueryScan:
outerChild = ((SubqueryScanState*)planstate)->subplan;
if (outerChild != NULL) {
if (datanode) {
outer_smp_idx = outerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, outerChild->plan->plan_node_id, outer_smp_idx);
} else
tmp_instr = planstate->instrument;
if (tmp_instr != NULL) {
*outerCycles = tmp_instr->cpuusage.m_cycles;
*outterRows = tmp_instr->ntuples;
}
}
break;
case T_RemoteQuery:
case T_VecRemoteQuery:
case T_Stream:
case T_VecStream:
break;
default:
outerChild = outerPlanState(planstate);
if (outerChild != NULL) {
if (datanode) {
outer_smp_idx = outerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, outerChild->plan->plan_node_id, outer_smp_idx);
} else
tmp_instr = outerChild->instrument;
if (tmp_instr != NULL) {
*outerCycles = tmp_instr->cpuusage.m_cycles;
*outterRows = tmp_instr->ntuples;
}
}
innerChild = innerPlanState(planstate);
if (innerChild != NULL) {
if (datanode) {
inner_smp_idx = innerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, innerChild->plan->plan_node_id, inner_smp_idx);
} else
tmp_instr = innerChild->instrument;
if (tmp_instr != NULL) {
*innerCycles = tmp_instr->cpuusage.m_cycles;
*innerRows = tmp_instr->ntuples;
}
}
break;
}
}
* Calculate the cpu cycles of a ModifyTable, Append, MergeAppend,
* BitmapAnd, or BitmapOr node.
*
* The ancestors list should already contain the total cpu cycles
* of the operator.
*/
template <bool datanode>
static void CalCPUMemberNode(const List* plans, PlanState** planstates, int idx, int smpIdx, double* Cycles, uint64* proRows)
{
int nplans = list_length(plans);
int k;
Instrumentation* tmp_instr = NULL;
for (k = 0; k < nplans; k++) {
if (datanode) {
int child_smp_idx = planstates[k]->plan->parallel_enabled ? smpIdx : 0;
tmp_instr =
u_sess->instr_cxt.global_instr->getInstrSlot(idx, planstates[k]->plan->plan_node_id, child_smp_idx);
} else {
tmp_instr = planstates[k]->instrument;
}
if (tmp_instr != NULL) {
*Cycles += tmp_instr->cpuusage.m_cycles;
*proRows += tmp_instr->ntuples;
}
}
}
template <bool datanode>
static void CalOperTimeMemberNode(const List* plans, PlanState** planstates, int idx, int smpIdx, double* oper_time)
{
int nplans = list_length(plans);
int k;
Instrumentation* tmp_instr = NULL;
for (k = 0; k < nplans; k++) {
if (datanode) {
int child_smp_idx = planstates[k]->plan->parallel_enabled ? smpIdx : 0;
tmp_instr =
u_sess->instr_cxt.global_instr->getInstrSlot(idx, planstates[k]->plan->plan_node_id, child_smp_idx);
} else {
tmp_instr = planstates[k]->instrument;
}
if (tmp_instr != NULL && tmp_instr->nloops > 0) {
*oper_time += 1000.0 * tmp_instr->total;
}
}
}
* Calculate the child node's run time
*/
template <bool datanode>
static void show_child_time(
PlanState* planstate, int idx, int smpIdx, double* inner_time, double* outer_time, ExplainState* es)
{
PlanState* outerChild = NULL;
PlanState* innerChild = NULL;
Instrumentation* tmp_instr = NULL;
Plan* plan = planstate->plan;
int outer_smp_idx = 0;
int inner_smp_idx = 0;
switch (nodeTag(plan)) {
case T_ModifyTable:
case T_VecModifyTable:
CalOperTimeMemberNode<datanode>(
((ModifyTable*)plan)->plans, ((ModifyTableState*)planstate)->mt_plans, idx, smpIdx, outer_time);
break;
case T_VecAppend:
case T_Append:
CalOperTimeMemberNode<datanode>(
((Append*)plan)->appendplans, ((AppendState*)planstate)->appendplans, idx, smpIdx, outer_time);
break;
case T_MergeAppend:
CalOperTimeMemberNode<datanode>(
((MergeAppend*)plan)->mergeplans, ((MergeAppendState*)planstate)->mergeplans, idx, smpIdx, outer_time);
break;
case T_BitmapAnd:
CalOperTimeMemberNode<datanode>(
((BitmapAnd*)plan)->bitmapplans, ((BitmapAndState*)planstate)->bitmapplans, idx, smpIdx, outer_time);
break;
case T_BitmapOr:
CalOperTimeMemberNode<datanode>(
((BitmapOr*)plan)->bitmapplans, ((BitmapOrState*)planstate)->bitmapplans, idx, smpIdx, outer_time);
break;
case T_CStoreIndexAnd:
CalOperTimeMemberNode<datanode>(((CStoreIndexAnd*)plan)->bitmapplans,
((BitmapAndState*)planstate)->bitmapplans,
idx,
smpIdx,
outer_time);
break;
case T_CStoreIndexOr:
CalOperTimeMemberNode<datanode>(
((CStoreIndexOr*)plan)->bitmapplans, ((BitmapOrState*)planstate)->bitmapplans, idx, smpIdx, outer_time);
break;
case T_SubqueryScan:
case T_VecSubqueryScan:
outerChild = ((SubqueryScanState*)planstate)->subplan;
if (outerChild != NULL) {
if (datanode) {
outer_smp_idx = outerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, outerChild->plan->plan_node_id, outer_smp_idx);
} else
tmp_instr = planstate->instrument;
if (tmp_instr != NULL && tmp_instr->nloops > 0)
*outer_time = 1000.0 * tmp_instr->total;
}
break;
case T_RemoteQuery:
case T_VecRemoteQuery:
*outer_time = planstate->instrument->network_perfdata.total_poll_time;
if (es->planinfo->m_query_summary) {
double deserialze_time = planstate->instrument->network_perfdata.total_deserialize_time;
appendStringInfo(
es->planinfo->m_query_summary->info_str, "Remote query poll time: %.3f ms", *outer_time);
appendStringInfo(
es->planinfo->m_query_summary->info_str, ", Deserialze time: %.3f ms\n", deserialze_time);
}
break;
case T_Stream:
case T_VecStream:
if (datanode) {
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(idx, planstate->plan->plan_node_id, smpIdx);
if (tmp_instr != NULL)
*outer_time = tmp_instr->network_perfdata.total_poll_time;
}
break;
default:
outerChild = outerPlanState(planstate);
if (outerChild != NULL) {
if (datanode) {
outer_smp_idx = outerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, outerChild->plan->plan_node_id, outer_smp_idx);
} else {
InstrEndLoop(outerChild->instrument);
tmp_instr = outerChild->instrument;
}
if (tmp_instr != NULL && tmp_instr->nloops > 0) {
*outer_time = 1000.0 * tmp_instr->total;
}
}
innerChild = innerPlanState(planstate);
if (innerChild != NULL) {
if (datanode) {
inner_smp_idx = innerChild->plan->parallel_enabled ? smpIdx : 0;
tmp_instr = u_sess->instr_cxt.global_instr->getInstrSlot(
idx, innerChild->plan->plan_node_id, inner_smp_idx);
} else {
InstrEndLoop(innerChild->instrument);
tmp_instr = innerChild->instrument;
}
if (tmp_instr != NULL && tmp_instr->nloops > 0) {
*inner_time = 1000.0 * tmp_instr->total;
}
}
break;
}
}
inline static void show_cpu_info(StringInfo infostr, double incCycles, double exCycles, uint64 proRows)
{
appendStringInfoString(infostr, "(CPU:");
appendStringInfo(infostr, " ex c/r=%.0f,", proRows != 0 ? (long)(exCycles / proRows) : 0.0);
appendStringInfo(infostr, " ex row=%lu,", proRows);
appendStringInfo(infostr, " ex cyc=%.0f,", exCycles);
appendStringInfo(infostr, " inc cyc=%.0f)", incCycles);
appendStringInfoChar(infostr, '\n');
}
static void show_detail_cpu(ExplainState* es, PlanState* planstate)
{
Instrumentation* instr = NULL;
uint64 proRows = 0;
double incCycles = 0.0;
double exCycles = 0.0;
double incCycles_max = -1;
double outerCycles = 0.0;
double innerCycles = 0.0;
uint64 outterRows = 0;
uint64 innerRows = 0;
double incCycles_min = HUGE_VAL - 1;
double exCycles_max = -1;
double exCycles_min = HUGE_VAL - 1;
double proRows_max = -1;
double proRows_min = HUGE_VAL - 1;
bool is_null = false;
int i = 0;
int j = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (u_sess->instr_cxt.global_instr->getInstruNodeNum() > 0) {
for (i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
outerCycles = 0.0;
innerCycles = 0.0;
if (instr != NULL && instr->nloops > 0) {
if (!is_null)
is_null = true;
show_child_cpu_cycles_and_rows<true>(
planstate, i, j, &outerCycles, &innerCycles, &outterRows, &innerRows);
incCycles = instr->cpuusage.m_cycles;
exCycles = incCycles - outerCycles - innerCycles;
proRows = (long)(instr->ntuples / instr->nloops);
proRows = (proRows != 0 ? (long)(exCycles / proRows) : 0.0);
incCycles_max = rtl::max(incCycles_max, incCycles);
incCycles_min = rtl::min(incCycles_min, incCycles);
exCycles_max = rtl::max(exCycles_max, exCycles);
exCycles_min = rtl::min(exCycles_min, exCycles);
proRows_max = rtl::max((uint64)proRows_max, proRows);
proRows_min = rtl::min((uint64)proRows_min, proRows);
}
}
}
}
if (is_null) {
AssertEreport(
(incCycles_max != -1 || exCycles_max != -1 || proRows_max != -1) &&
(incCycles_min != HUGE_VAL - 1 || exCycles_min != HUGE_VAL - 1 || proRows_min != HUGE_VAL - 1),
MOD_EXECUTOR,
"unexpect min/max value");
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, "CPU:");
appendStringInfo(es->str, " max_ex c/r=%.0f,", proRows_max);
appendStringInfo(es->str, " max_ex cyc=%.0f,", exCycles_max);
appendStringInfo(es->str, " max_inc cyc=%.0f", incCycles_max);
appendStringInfoChar(es->str, '\n');
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, " ");
appendStringInfo(es->str, " min_ex c/r=%.0f,", proRows_min);
appendStringInfo(es->str, " min_ex cyc=%.0f,", exCycles_min);
appendStringInfo(es->str, " min_inc cyc=%.0f", incCycles_min);
appendStringInfoChar(es->str, '\n');
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_IOInfo) {
es->planinfo->m_IOInfo->set_plan_name<true, true>();
appendStringInfoString(es->planinfo->m_IOInfo->info_str, "CPU:");
appendStringInfo(es->planinfo->m_IOInfo->info_str, " max_ex c/r=%.0f,", proRows_max);
appendStringInfo(es->planinfo->m_IOInfo->info_str, " max_ex cyc=%.0f,", exCycles_max);
appendStringInfo(es->planinfo->m_IOInfo->info_str, " max_inc cyc=%.0f", incCycles_max);
appendStringInfoChar(es->planinfo->m_IOInfo->info_str, '\n');
appendStringInfoSpaces(es->planinfo->m_IOInfo->info_str, 11);
appendStringInfo(es->planinfo->m_IOInfo->info_str, " min_ex c/r=%.0f,", proRows_min);
appendStringInfo(es->planinfo->m_IOInfo->info_str, " min_ex cyc=%.0f,", exCycles_min);
appendStringInfo(es->planinfo->m_IOInfo->info_str, " min_inc cyc=%.0f", incCycles_min);
appendStringInfoChar(es->planinfo->m_IOInfo->info_str, '\n');
}
} else {
ExplainPropertyLong("Exclusive Max Cycles Per Row", (long)proRows_max, es);
ExplainPropertyLong("Exclusive Max Cycles", (long)exCycles_max, es);
ExplainPropertyLong("Inclusive Max Cycles", (long)incCycles_max, es);
ExplainPropertyLong("Exclusive Min Cycles Per Row", (long)proRows_min, es);
ExplainPropertyLong("Exclusive Min Cycles", (long)exCycles_min, es);
ExplainPropertyLong("Inclusive Min Cycles", (long)incCycles_min, es);
}
} else if (es->analyze) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, "(CPU: never executed)");
appendStringInfoChar(es->str, '\n');
} else {
ExplainPropertyFloat("Actual Exclusive Per Row", 0.0, 0, es);
ExplainPropertyFloat("Actual Exclusive Cycles", 0.0, 0, es);
ExplainPropertyFloat("Actual Inclusive Cycles", 0.0, 0, es);
}
}
}
static bool get_execute_mode(const ExplainState* es, int idx)
{
int id;
if (!es->from_dn || !es->sql_execute)
return true;
id = idx < 0 ? 0 : idx;
if (u_sess->instr_cxt.global_instr) {
ThreadInstrumentation* threadinstr =
#ifdef ENABLE_MULTIPLE_NODES
u_sess->instr_cxt.global_instr->getThreadInstrumentationCN(id);
#else
u_sess->instr_cxt.global_instr->getThreadInstrumentationDN(1, 0);
#endif
if (threadinstr == NULL)
return true;
bool execute = threadinstr->m_instrArray[0].instr.isExecute;
if (!execute) {
if (es->datanodeinfo.all_datanodes)
return false;
for (int i = 0; i < es->datanodeinfo.len_nodelist; i++) {
if (id == es->datanodeinfo.node_index[i])
return false;
}
}
}
return true;
}
static void show_cpu(ExplainState* es, const Instrumentation* instrument, double innerCycles, double outerCycles, int nodeIdx,
int smpIdx, uint64 proRows)
{
if (instrument != NULL && instrument->nloops > 0) {
double incCycles = 0.0;
double exCycles = 0.0;
int64 ex_cyc_rows;
incCycles = instrument->cpuusage.m_cycles;
exCycles = incCycles - outerCycles - innerCycles;
ex_cyc_rows = proRows != 0 ? (long)(exCycles / proRows) : 0;
if (es->planinfo && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, EX_CYC, Int64GetDatum(exCycles));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, INC_CYC, Int64GetDatum(incCycles));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, EX_CYC_PER_ROWS, Int64GetDatum(ex_cyc_rows));
show_cpu_info(es->planinfo->m_datanodeInfo->info_str, incCycles, exCycles, proRows);
return;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
show_cpu_info(es->str, incCycles, exCycles, proRows);
} else {
ExplainPropertyLong("Exclusive Cycles Per Row", proRows != 0 ? (long)(exCycles / proRows) : 0, es);
ExplainPropertyLong("Exclusive Cycles", (long)exCycles, es);
ExplainPropertyLong("Inclusive Cycles", (long)incCycles, es);
}
} else if (es->analyze) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (get_execute_mode(es, nodeIdx))
appendStringInfo(es->str, "(CPU: never executed)");
else
appendStringInfo(es->str, "(CPU: unknown)");
appendStringInfoChar(es->str, '\n');
} else {
ExplainPropertyFloat("Actual Exclusive Cycles Per Row", 0.0, 0, es);
ExplainPropertyFloat("Actual Exclusive Cycles", 0.0, 0, es);
ExplainPropertyFloat("Actual Inclusive Cycles", 0.0, 0, es);
}
}
}
* @Description: show memory info about each plan node in pretty mode
* @in es - the explain state info
* @in node_name - datanode name
* @in instrument - instrument info
* @in smpIdx - current smp idx
* @in dop - query dop
* @out - void
*/
static void show_pretty_memory(ExplainState* es, char* node_name, const Instrumentation* instrument, int smpIdx, int dop)
{
es->planinfo->m_staticInfo->set_plan_name<true, true>();
if (node_name != NULL) {
es->planinfo->m_staticInfo->set_datanode_name(node_name, smpIdx, dop);
}
if (instrument->memoryinfo.peakOpMemory <= 1024)
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Peak Memory: %ldBYTE", instrument->memoryinfo.peakOpMemory);
else
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "Peak Memory: %ldKB", instrument->memoryinfo.peakOpMemory / 1024);
if (instrument->memoryinfo.peakControlMemory > 0) {
if (instrument->memoryinfo.peakControlMemory <= 1024)
appendStringInfo(es->planinfo->m_staticInfo->info_str,
", Control Memory: %ldBYTE",
instrument->memoryinfo.peakControlMemory);
else
appendStringInfo(es->planinfo->m_staticInfo->info_str,
", Control Memory: %ldKB",
instrument->memoryinfo.peakControlMemory / 1024);
}
if (instrument->memoryinfo.operatorMemory > 0) {
if (instrument->memoryinfo.operatorMemory <= 1024)
appendStringInfo(
es->planinfo->m_staticInfo->info_str, ", Estimate Memory: %dKB", instrument->memoryinfo.operatorMemory);
else
appendStringInfo(es->planinfo->m_staticInfo->info_str,
", Estimate Memory: %dMB",
instrument->memoryinfo.operatorMemory / 1024);
}
if (instrument->spreadNum > 0) {
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", Auto spread times: %d", instrument->spreadNum);
}
if (instrument->sysBusy) {
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", Early spilled");
}
if (instrument->width > 0) {
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", Width: %d", instrument->width);
}
appendStringInfo(es->planinfo->m_staticInfo->info_str, "\n");
}
static void show_pretty_time(
ExplainState* es, Instrumentation* instrument, char* node_name, int nodeIdx, int smpIdx, int dop, bool executed)
{
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double startup_sec = 1000.0 * instrument->startup;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
if (es->format == EXPLAIN_FORMAT_TEXT) {
char str[100] = "\0";
int rc = sprintf_s(str, sizeof(str), "%.3f", total_sec);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_ROWS, Float8GetDatum(rows));
if (es->timing) {
es->planinfo->m_planInfo->put(ACTUAL_TIME, PointerGetDatum(cstring_to_text(str)));
}
char memoryPeakSize[100] = "\0";
if (instrument->memoryinfo.peakOpMemory <= 1024)
rc = sprintf_s(memoryPeakSize, sizeof(memoryPeakSize), "%ldBYTE", instrument->memoryinfo.peakOpMemory);
else
rc = sprintf_s(
memoryPeakSize, sizeof(memoryPeakSize), "%ldKB", instrument->memoryinfo.peakOpMemory / 1024);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_WIDTH, PointerGetDatum(cstring_to_text("")));
es->planinfo->m_planInfo->put(ACTUAL_MEMORY, PointerGetDatum(cstring_to_text(memoryPeakSize)));
if (es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, START_TIME, Float8GetDatum(startup_sec));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, TOTAL_TIME, Float8GetDatum(total_sec));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, PLAN_ROWS, rows);
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, PLAN_LOOPS, nloops);
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, PEAK_OP_MEMORY, instrument->memoryinfo.peakOpMemory);
if (node_name != NULL)
es->planinfo->m_runtimeinfo->put(
nodeIdx, smpIdx, NODE_NAME, PointerGetDatum(cstring_to_text(node_name)));
else
es->planinfo->m_runtimeinfo->put(
nodeIdx, smpIdx, NODE_NAME, PointerGetDatum(cstring_to_text("coordinator")));
}
if (es->detail) {
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
if (node_name != NULL) {
es->planinfo->m_datanodeInfo->set_datanode_name(node_name, smpIdx, dop);
}
if (instrument->need_timer)
appendStringInfo(es->planinfo->m_datanodeInfo->info_str,
"(actual time=%.3f..%.3f rows=%.0f loops=%.0f)\n",
startup_sec,
total_sec,
rows,
nloops);
else
appendStringInfo(
es->planinfo->m_datanodeInfo->info_str, "(actual rows=%.0f loops=%.0f)\n", rows, nloops);
show_pretty_memory(es, node_name, instrument, smpIdx, dop);
}
}
} else if (es->planinfo->m_runtimeinfo) {
if (node_name != NULL) {
if (executed)
es->planinfo->m_runtimeinfo->put(
nodeIdx, smpIdx, NODE_NAME, PointerGetDatum(cstring_to_text(node_name)));
} else
es->planinfo->m_runtimeinfo->put(
nodeIdx, smpIdx, NODE_NAME, PointerGetDatum(cstring_to_text("coordinator")));
}
}
template <bool isdetail>
static StreamTime* get_instrument(
int j, int smpid, Track* trackpoint, int plan_node_id, StringInfo str, bool* first_time)
{
StreamTime* instrument = NULL;
int dop = u_sess->instr_cxt.global_instr->getNodeDop(plan_node_id);
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(j, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", j);
}
#endif
instrument = &trackpoint->track_time;
if (*first_time) {
if (trackpoint->node_id > 0)
appendStringInfo(str,
"Plan Node id: %d Track name: %s\n",
trackpoint->node_id,
trackdesc[trackpoint->registerTrackId].trackName);
if (trackpoint->node_id == -1)
appendStringInfo(str,
"Segment Id: %d Track name: %s\n",
plan_node_id,
trackdesc[trackpoint->registerTrackId].trackName);
*first_time = false;
}
if (isdetail) {
appendStringInfoSpaces(str, 6);
if (dop == 1)
appendStringInfo(str, " %s", node_name);
else
appendStringInfo(str, "%s[worker %d]", node_name, smpid);
}
StreamEndLoop(instrument);
return instrument;
}
static void show_track_time_without_plannodeid(ExplainState* es)
{
StringInfo str;
StreamTime* instrument = NULL;
int i = 0;
int j = 0;
bool has_perf = CPUMon::m_has_perf;
AccumCounters accumcount;
const char* track_name = NULL;
bool first_time = true;
int track_Id;
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL)
str = es->str;
else
str = es->planinfo->m_profileInfo->info_str;
int m_num_streams = u_sess->instr_cxt.global_instr->getInstruThreadNum();
int m_query_dop = u_sess->instr_cxt.global_instr->get_query_dop();
int threadlen = u_sess->instr_cxt.global_instr->get_threadInstrArrayLen();
int m_gather_count = u_sess->instr_cxt.global_instr->get_gather_num();
ThreadInstrumentation* threadinstr = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(0);
int generaltracknum = threadinstr->get_generaltracknum();
for (i = 0; i < generaltracknum; i++) {
track_Id = i;
track_name = trackdesc[track_Id].trackName;
ThreadInstrumentation* threadinstr_tmp = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(0);
Track* generaltrack = &threadinstr_tmp->m_generalTrackArray[i];
int segmentid = threadinstr_tmp->getSegmentId();
if (segmentid > 0 && generaltrack->registerTrackId == track_Id && IS_NULL(generaltrack->active)) {
instrument = &generaltrack->track_time;
StreamEndLoop(instrument);
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
appendStringInfo(str, "Segment Id: %d Track name: %s\n", segmentid, track_name);
#ifdef ENABLE_MULTIPLE_NODES
char* nodename = PGXCNodeGetNodeNameFromId(0, PGXC_NODE_COORDINATOR);
#else
char* nodename = NULL;
if (!IS_SPQ_RUNNING) {
nodename = g_instance.exec_cxt.nodeName;
} else {
nodename = (char*)palloc0(SPQNODENAMELEN);
sprintf(nodename, "%d", i);
}
#endif
appendStringInfoSpaces(str, 6);
appendStringInfo(str, " %s:", nodename);
if (instrument->need_timer) {
if (!has_perf) {
appendStringInfo(str, " (time=%.3f total_calls=%.0f loops=%.0f)", total_sec, rows, nloops);
} else {
accumcount = generaltrack->accumCounters;
appendStringInfo(str,
" (time=%.3f Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
total_sec,
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
} else {
if (!has_perf) {
appendStringInfo(str, " (total_calls=%.0f loops=%.0f)", rows, nloops);
} else {
accumcount = generaltrack->accumCounters;
appendStringInfo(str,
" (Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
}
appendStringInfoChar(str, '\n');
}
}
for (j = 1; j < threadlen; j++) {
ThreadInstrumentation* threadinstr_tmp = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(j);
if (threadinstr_tmp != NULL) {
Track* generaltrack_tmp = &threadinstr_tmp->m_generalTrackArray[i];
int segmentid_tmp = threadinstr_tmp->getSegmentId();
if (segmentid_tmp > 0 && generaltrack_tmp->registerTrackId == track_Id &&
IS_NULL(generaltrack_tmp->active)) {
int nodeidx = (j - 1) / ((m_gather_count + m_num_streams) * m_query_dop);
int smpid = (j - 1) % m_query_dop;
instrument = get_instrument<true>(nodeidx, smpid, generaltrack_tmp, segmentid_tmp, str,
&first_time);
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
if (instrument->need_timer) {
if (!has_perf) {
appendStringInfo(
str, " (time=%.3f total_calls=%.0f loops=%.0f)", total_sec, rows, nloops);
} else {
accumcount = generaltrack_tmp->accumCounters;
appendStringInfo(str,
" (time=%.3f Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
total_sec,
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
} else {
if (!has_perf) {
appendStringInfo(str, " (total_calls=%.0f loops=%.0f)", rows, nloops);
} else {
accumcount = generaltrack_tmp->accumCounters;
appendStringInfo(str,
" (Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
}
appendStringInfoChar(str, '\n');
}
}
}
}
}
}
static void show_track_time_info(ExplainState* es)
{
StreamTime* instrument = NULL;
Track* trackArray = NULL;
bool first_time = true;
int i = 0;
int j = 0;
StringInfo str;
int plan_node_id;
AccumCounters accumcount;
const char* track_name = NULL;
bool has_perf = CPUMon::m_has_perf;
int track_Id;
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL) {
str = es->str;
appendStringInfo(str, "*********User Define Profiling*********\n");
} else
str = es->planinfo->m_profileInfo->info_str;
int m_num_streams = u_sess->instr_cxt.global_instr->getInstruThreadNum();
int m_query_dop = u_sess->instr_cxt.global_instr->get_query_dop();
int plannodes_num = u_sess->instr_cxt.global_instr->getInstruPlanSize() + 1;
int m_gather_count = u_sess->instr_cxt.global_instr->get_gather_num();
ThreadInstrumentation* threadinstr = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(0);
int tracknum = threadinstr->get_tracknum();
int generaltracknum = threadinstr->get_generaltracknum();
int dn_start_node_id = u_sess->instr_cxt.global_instr->get_startnodeid();
if (es->detail) {
show_track_time_without_plannodeid(es);
for (i = 0; i < tracknum; i++) {
track_Id = i + generaltracknum;
track_name = trackdesc[track_Id].trackName;
ThreadInstrumentation* threadinstr_tmp = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(0);
int trackArrayLen = threadinstr_tmp->get_instrArrayLen();
for (j = 0; j < trackArrayLen; j++) {
trackArray = threadinstr_tmp->m_instrArray[j].tracks;
Track* trackpoint = &trackArray[i];
plan_node_id = trackpoint->node_id;
if (plan_node_id > 0 && trackpoint->registerTrackId == track_Id && IS_NULL(trackpoint->active)) {
instrument = &trackpoint->track_time;
StreamEndLoop(instrument);
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
appendStringInfo(str, "Plan Node id: %d Track name: %s\n", plan_node_id, track_name);
#ifdef ENABLE_MULTIPLE_NODES
char* nodename = PGXCNodeGetNodeNameFromId(0, PGXC_NODE_COORDINATOR);
#else
char* nodename = NULL;
if (!IS_SPQ_RUNNING) {
nodename = g_instance.exec_cxt.nodeName;
} else {
nodename = (char*)palloc0(SPQNODENAMELEN);
sprintf(nodename, "%d", i);
}
#endif
appendStringInfoSpaces(str, 6);
appendStringInfo(str, " %s:", nodename);
if (instrument->need_timer) {
if (!has_perf) {
appendStringInfo(
str, " (time=%.3f total_calls=%.0f loops=%.0f)", total_sec, rows, nloops);
} else {
accumcount = trackpoint->accumCounters;
appendStringInfo(str,
" (time=%.3f Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
total_sec,
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
} else {
if (!has_perf) {
appendStringInfo(str, " (total_calls=%.0f loops=%.0f)", rows, nloops);
} else {
accumcount = trackpoint->accumCounters;
appendStringInfo(str,
" (Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
}
appendStringInfoChar(str, '\n');
}
}
}
}
for (j = dn_start_node_id; j < plannodes_num; j++) {
for (i = 0; i < tracknum; i++) {
first_time = true;
track_Id = i + generaltracknum;
for (int k = 0; k < u_sess->instr_cxt.global_instr->getInstruNodeNum(); k++) {
for (int m = 0; m < u_sess->instr_cxt.global_instr->getNodeDop(j); m++) {
ThreadInstrumentation* threadinstr_tmp =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(k, j, m);
if (threadinstr_tmp != NULL) {
trackArray = threadinstr_tmp->get_tracks(j);
if (trackArray == NULL)
continue;
Track* trackpoint = &trackArray[i];
plan_node_id = trackpoint->node_id;
if (plan_node_id > 0 && trackpoint->registerTrackId == track_Id &&
IS_NULL(trackpoint->active)) {
instrument = get_instrument<true>(k, m, trackpoint, plan_node_id, str, &first_time);
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
if (instrument->need_timer) {
if (!has_perf)
appendStringInfo(str,
" (time=%.3f total_calls=%.0f loops=%.0f)",
total_sec,
rows,
nloops);
else {
accumcount = trackpoint->accumCounters;
appendStringInfo(str,
" (time=%.3f Instructions=%ld cpu cycles=%ld total_calls=%.0f "
"loops=%.0f)",
total_sec,
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
} else {
if (!has_perf)
appendStringInfo(str, " (total_calls=%.0f loops=%.0f)", rows, nloops);
else {
accumcount = trackpoint->accumCounters;
appendStringInfo(str,
" (Instructions=%ld cpu cycles=%ld total_calls=%.0f loops=%.0f)",
accumcount.accumCounters[1],
accumcount.accumCounters[0],
rows,
nloops);
}
}
appendStringInfoChar(str, '\n');
}
}
}
}
}
}
}
} else {
for (i = 0; i < generaltracknum; i++) {
double total_sec_max = -1;
double total_sec_min = HUGE_VAL - 1;
bool need_timer = false;
double max_calls = -1;
double min_calls = HUGE_VAL - 1;
double total_sec;
double calls = 0;
int64 max_stru = -1;
int64 min_stru = LONG_MAX;
int64 max_cpu_cycle = -1;
int64 min_cpu_cycle = LONG_MAX;
track_Id = i;
first_time = true;
for (j = 0; j < u_sess->instr_cxt.global_instr->get_threadInstrArrayLen(); j++) {
ThreadInstrumentation* threadinstr_tmp = u_sess->instr_cxt.global_instr->get_cnthreadinstrumentation(j);
if (threadinstr_tmp != NULL) {
Track* trackpoint = &threadinstr_tmp->m_generalTrackArray[i];
int segmentid = threadinstr_tmp->getSegmentId();
if (segmentid > 0 && trackpoint->registerTrackId == track_Id && IS_NULL(trackpoint->active)) {
int nodeidx = (j == 0 ? 0 : (j - 1) / ((m_gather_count + m_num_streams) * m_query_dop));
int smpid = (j == 0 ? 0 : (j - 1) % m_query_dop);
instrument = get_instrument<false>(nodeidx, smpid, trackpoint, segmentid, str, &first_time);
if (instrument != NULL && instrument->nloops > 0) {
if (!need_timer)
need_timer = instrument->need_timer;
calls = instrument->ntuples;
max_calls = rtl::max(max_calls, calls);
min_calls = rtl::min(min_calls, calls);
total_sec = 1000.0 * instrument->total;
total_sec_max = rtl::max(total_sec_max, total_sec);
total_sec_min = rtl::min(total_sec_min, total_sec);
}
if (has_perf) {
accumcount = trackpoint->accumCounters;
max_stru = rtl::max(max_stru, accumcount.accumCounters[1]);
min_stru = rtl::min(min_stru, accumcount.accumCounters[1]);
max_cpu_cycle = rtl::max(max_cpu_cycle, accumcount.accumCounters[0]);
min_cpu_cycle = rtl::min(min_cpu_cycle, accumcount.accumCounters[0]);
}
}
}
}
if (need_timer && es->format == EXPLAIN_FORMAT_TEXT) {
AssertEreport(
total_sec_max != -1 && total_sec_min != HUGE_VAL - 1, MOD_EXECUTOR, "unexpect min/max value");
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (!has_perf)
appendStringInfo(str,
" (actual time=[%.3f, %.3f], calls=[%.0f, %.0f])\n",
total_sec_min,
total_sec_max,
min_calls,
max_calls);
else
appendStringInfo(str,
" (actual time=[%.3f, %.3f], Instructions=[%ld, %ld], cpu cycles=[%ld, %ld], calls=[%.0f, "
"%.0f])\n",
total_sec_min,
total_sec_max,
min_stru,
max_stru,
min_cpu_cycle,
max_cpu_cycle,
min_calls,
max_calls);
}
}
}
for (j = 1; j < plannodes_num; j++) {
for (i = 0; i < tracknum; i++) {
double total_sec_max = -1;
double total_sec_min = HUGE_VAL - 1;
bool need_timer = false;
double max_calls = -1;
double min_calls = HUGE_VAL - 1;
double total_sec;
double calls = 0;
int64 max_stru = -1;
int64 min_stru = LONG_MAX;
int64 max_cpu_cycle = -1;
int64 min_cpu_cycle = LONG_MAX;
first_time = true;
track_Id = i + generaltracknum;
for (int k = 0; k < u_sess->instr_cxt.global_instr->getInstruNodeNum(); k++) {
for (int m = 0; m < u_sess->instr_cxt.global_instr->getNodeDop(j); m++) {
ThreadInstrumentation* threadinstr_tmp =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(k, j, m);
if (threadinstr_tmp != NULL) {
Track* trackArray_tmp = threadinstr_tmp->get_tracks(j);
if (trackArray_tmp == NULL)
continue;
Track* trackpoint = &trackArray_tmp[i];
plan_node_id = trackpoint->node_id;
if (plan_node_id > 0 && trackpoint->registerTrackId == track_Id &&
IS_NULL(trackpoint->active)) {
instrument = get_instrument<false>(k, m, trackpoint, plan_node_id, str, &first_time);
if (instrument != NULL && instrument->nloops > 0) {
if (!need_timer)
need_timer = instrument->need_timer;
calls = instrument->ntuples;
max_calls = rtl::max(max_calls, calls);
min_calls = rtl::min(min_calls, calls);
total_sec = 1000.0 * instrument->total;
total_sec_max = rtl::max(total_sec_max, total_sec);
total_sec_min = rtl::min(total_sec_min, total_sec);
}
if (has_perf) {
accumcount = trackpoint->accumCounters;
max_stru = rtl::max(max_stru, accumcount.accumCounters[1]);
min_stru = rtl::min(min_stru, accumcount.accumCounters[1]);
max_cpu_cycle = rtl::max(max_cpu_cycle, accumcount.accumCounters[0]);
min_cpu_cycle = rtl::min(min_cpu_cycle, accumcount.accumCounters[0]);
}
}
}
}
}
if (need_timer && es->format == EXPLAIN_FORMAT_TEXT) {
AssertEreport(
total_sec_max != -1 && total_sec_min != HUGE_VAL - 1, MOD_EXECUTOR, "unexpect min/max value");
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (!has_perf)
appendStringInfo(str,
" (actual time=[%.3f, %.3f], calls=[%.0f, %.0f])\n",
total_sec_min,
total_sec_max,
min_calls,
max_calls);
else
appendStringInfo(str,
" (actual time=[%.3f, %.3f], Instructions=[%ld, %ld], cpu cycles=[%ld, %ld], "
"calls=[%.0f, %.0f])\n",
total_sec_min,
total_sec_max,
min_stru,
max_stru,
min_cpu_cycle,
max_cpu_cycle,
min_calls,
max_calls);
}
}
}
}
}
}
int get_track_time(ExplainState* es, PlanState* planstate, bool show_track, bool show_buffer, bool show_dummygroup,
bool show_indexinfo, bool show_storageinfo)
{
if (show_track)
show_track_time_info(es);
if (show_buffer) {
append_datanode_name(es, "datanode1", 2, 1);
show_analyze_buffers(es, planstate, es->str, 1);
show_buffers(es, es->str, u_sess->instr_cxt.global_instr->getInstrSlot(0, 1), false, 0, 0, NULL);
}
if (show_dummygroup) {
ExplainDummyGroup("ut", "test", es);
ExplainDummyGroup("ut", NULL, es);
}
if (show_indexinfo) {
ExplainIndexScanDetails(0, BackwardScanDirection, es);
ExplainIndexScanDetails(0, NoMovementScanDirection, es);
ExplainIndexScanDetails(0, ForwardScanDirection, es);
}
if (show_storageinfo) {
append_datanode_name(es, "datanode1", 2, 1);
show_detail_storage_info_json(u_sess->instr_cxt.global_instr->getInstrSlot(0, 1), es->str, es);
}
return 0;
}
template <bool is_detail>
static void show_time(ExplainState* es, const Instrumentation* instrument, int idx)
{
if (instrument != NULL && instrument->nloops > 0) {
double nloops = instrument->nloops;
const double startup_sec = 1000.0 * instrument->startup;
const double total_sec = 1000.0 * instrument->total;
double rows = instrument->ntuples;
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_detail == false)
appendStringInfoSpaces(es->str, 1);
if (instrument->need_timer)
appendStringInfo(
es->str, "(actual time=%.3f..%.3f rows=%.0f loops=%.0f)", startup_sec, total_sec, rows, nloops);
else
appendStringInfo(es->str, "(actual rows=%.0f loops=%.0f)", rows, nloops);
} else {
if (instrument->need_timer) {
ExplainPropertyFloat("Actual Startup Time", startup_sec, 3, es);
ExplainPropertyFloat("Actual Total Time", total_sec, 3, es);
}
ExplainPropertyFloat("Actual Rows", rows, 0, es);
ExplainPropertyFloat("Actual Loops", nloops, 0, es);
}
} else if (es->analyze) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (is_detail == false)
appendStringInfoSpaces(es->str, 1);
if (get_execute_mode(es, idx))
appendStringInfo(es->str, "(Actual time: never executed)");
else
appendStringInfo(es->str, "(Actual time: unknown)");
} else if (instrument != NULL && instrument->need_timer) {
ExplainPropertyFloat("Actual Startup Time", 0.0, 3, es);
ExplainPropertyFloat("Actual Total Time", 0.0, 3, es);
} else {
ExplainPropertyFloat("Actual Rows", 0.0, 0, es);
ExplainPropertyFloat("Actual Loops", 0.0, 0, es);
}
}
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoChar(es->str, '\n');
}
* @Description: get max plan id from a plannedstmt
* @in value -a integer
* @out int - digit of value
*/
static int get_digit(int value)
{
AssertEreport(value >= 0, MOD_EXECUTOR, "unexpect input value");
int result = 0;
while (value > 0) {
result++;
value /= 10;
}
if (result == 0) {
return 1;
}
return result;
}
template <bool datanode>
static void get_oper_time(ExplainState* es, PlanState* planstate, const Instrumentation* instr, int nodeIdx, int smpIdx)
{
double outer_total_time = 0.0;
double inner_total_time = 0.0;
double oper_time;
show_child_time<datanode>(planstate, nodeIdx, smpIdx, &inner_total_time, &outer_total_time, es);
if (instr != NULL && instr->nloops > 0) {
oper_time = 1000.0 * instr->total;
oper_time = oper_time - outer_total_time - inner_total_time;
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, OPER_TIME, Float8GetDatum(oper_time));
}
}
* @Description: show time info about stream send data to other thread
* @in es - the explain state info
* @in planstate - current planstate
* @out - void
*/
static void show_stream_send_time(ExplainState* es, const PlanState* planstate)
{
bool isSend = false;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL || es->detail == false)
return;
isSend = u_sess->instr_cxt.global_instr->IsSend(0, planstate->plan->plan_node_id, 0);
if (isSend == false) {
return;
}
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (int j = 0; j < dop; j++) {
Instrumentation* instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->stream_senddata.loops == true) {
double send_time = instr->stream_senddata.total_send_time;
double wait_quota_time = instr->stream_senddata.total_wait_quota_time;
double os_send_time = instr->stream_senddata.total_os_send_time;
double serialize_time = instr->stream_senddata.total_serialize_time;
double copy_time = instr->stream_senddata.total_copy_time;
es->planinfo->m_staticInfo->set_plan_name<false, true>();
es->planinfo->m_staticInfo->set_datanode_name(node_name, j, dop);
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Stream Send time: %.3f", send_time);
if (wait_quota_time > 0.0)
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", Wait Quota time: %.3f", wait_quota_time);
if (os_send_time > 0.0)
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", OS Kernel Send time: %.3f", os_send_time);
if (copy_time > 0.0)
appendStringInfo(es->planinfo->m_staticInfo->info_str, ", Data Copy time: %.3f", copy_time);
else
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "; Data Serialize time: %.3f", serialize_time);
appendStringInfoChar(es->planinfo->m_staticInfo->info_str, '\n');
}
}
}
}
static void show_datanode_time(ExplainState* es, PlanState* planstate)
{
Instrumentation* instr = NULL;
double total_sec_max = -1;
double total_sec_min = HUGE_VAL - 1;
double start_sec_max = -1;
double start_sec_min = HUGE_VAL - 1;
bool need_timer = false;
double rows = 0;
double total_sec;
double start_sec;
int64 peak_memory_min = (int64)(0x6FFFFFFFFFFFFFFF);
int64 peak_memory_max = 0;
int width_min = (int)(0x6FFFFFFF);
int width_max = 0;
bool executed = true;
int i = 0;
int j = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (es->detail) {
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoChar(es->str, '\n');
ExplainOpenGroup("Actual In Detail", "Actual In Detail", false, es);
for (i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
ThreadInstrumentation* threadinstr = u_sess->instr_cxt.global_instr->getThreadInstrumentationCN(i);
#else
char* node_name = NULL;
ThreadInstrumentation* threadinstr = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
threadinstr = u_sess->instr_cxt.global_instr->getThreadInstrumentationDN(1, 0);
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
threadinstr = u_sess->instr_cxt.global_instr->getThreadInstrumentationCN(i);
}
#endif
if (threadinstr == NULL)
continue;
executed = threadinstr->m_instrArray[0].instr.isExecute;
for (j = 0; j < dop; j++) {
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr == NULL)
continue;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
show_pretty_time(es, instr, node_name, i, j, dop, executed);
get_oper_time<true>(es, planstate, instr, i, j);
if (instr->needRCInfo) {
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
es->planinfo->m_datanodeInfo->set_datanode_name(node_name, j, dop);
ShowRoughCheckInfo(es, instr, i, j);
}
if (instr != NULL && instr->nloops > 0) {
if (!need_timer)
need_timer = instr->need_timer;
rows += instr->ntuples;
total_sec = 1000.0 * instr->total;
start_sec = 1000.0 * instr->startup;
total_sec_max = rtl::max(total_sec_max, total_sec);
total_sec_min = rtl::min(total_sec_min, total_sec);
start_sec_max = rtl::max(start_sec_max, start_sec);
start_sec_min = rtl::min(start_sec_min, start_sec);
peak_memory_min = rtl::min(peak_memory_min, instr->memoryinfo.peakOpMemory);
peak_memory_max = rtl::max(peak_memory_max, instr->memoryinfo.peakOpMemory);
width_min = rtl::min(width_min, instr->width);
width_max = rtl::max(width_max, instr->width);
}
} else {
append_datanode_name(es, node_name, dop, j);
show_time<true>(es, instr, i);
ShowRoughCheckInfo(es, instr, i, j);
}
ExplainCloseGroup("Plan", NULL, true, es);
}
}
ExplainCloseGroup("Actual In Detail", "Actual In Detail", false, es);
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->analyze) {
char str[100] = "\0";
int rc = 0;
if (need_timer)
rc = sprintf_s(str, sizeof(str), "[%.3f,%.3f]", total_sec_min, total_sec_max);
else
rc = sprintf_s(str, sizeof(str), "[0,0]");
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_ROWS, Float8GetDatum(rows));
if (es->timing) {
es->planinfo->m_planInfo->put(ACTUAL_TIME, PointerGetDatum(cstring_to_text(str)));
}
char width[50] = "\0";
if (width_max > 0) {
rc = sprintf_s(width, sizeof(width), "[%d,%d]", width_min, width_max);
securec_check_ss(rc, "\0", "\0");
}
es->planinfo->m_planInfo->put(ACTUAL_WIDTH, PointerGetDatum(cstring_to_text(width)));
if (es->planinfo->m_planInfo->m_query_mem_mode) {
char opmem[50] = "\0";
if (planstate->plan->operatorMemKB[0] > 0) {
if (planstate->plan->operatorMemKB[0] < 1024)
rc = sprintf_s(opmem, sizeof(opmem), "%dKB", (int)planstate->plan->operatorMemKB[0]);
else {
if (planstate->plan->operatorMemKB[0] > MIN_OP_MEM &&
planstate->plan->operatorMaxMem > planstate->plan->operatorMemKB[0])
rc = sprintf_s(opmem,
sizeof(opmem),
"%dMB(%dMB)",
(int)planstate->plan->operatorMemKB[0] / 1024,
(int)planstate->plan->operatorMaxMem / 1024);
else
rc = sprintf_s(
opmem, sizeof(opmem), "%dMB", (int)planstate->plan->operatorMemKB[0] / 1024);
}
securec_check_ss(rc, "\0", "\0");
}
es->planinfo->m_planInfo->put(ESTIMATE_MEMORY, PointerGetDatum(cstring_to_text(opmem)));
}
char memoryPeakMinSize[100] = "\0";
if (peak_memory_min < 1024)
rc = sprintf_s(memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldBYTE", peak_memory_min);
else if (peak_memory_min < 1024 * 1024)
rc = sprintf_s(memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldKB", peak_memory_min / 1024);
else
rc = sprintf_s(
memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldMB", peak_memory_min / (1024 * 1024));
securec_check_ss(rc, "\0", "\0");
char memoryPeakMaxSize[100] = "\0";
if (peak_memory_max < 1024)
rc = sprintf_s(memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldBYTE", peak_memory_max);
else if (peak_memory_max < 1024 * 1024)
rc = sprintf_s(memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldKB", peak_memory_max / 1024);
else
rc = sprintf_s(
memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldMB", peak_memory_max / (1024 * 1024));
securec_check_ss(rc, "\0", "\0");
char memoryStr[256] = "\0";
if (need_timer)
rc = sprintf_s(memoryStr, sizeof(memoryStr), "[%s, %s]", memoryPeakMinSize, memoryPeakMaxSize);
else
rc = sprintf_s(memoryStr, sizeof(memoryStr), "[0, 0]");
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_MEMORY, PointerGetDatum(cstring_to_text(memoryStr)));
}
}
} else {
if (u_sess->instr_cxt.global_instr->getInstruNodeNum() > 0) {
for (i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
if (!need_timer)
need_timer = true;
rows += instr->ntuples;
total_sec = 1000.0 * instr->total;
start_sec = 1000.0 * instr->startup;
total_sec_max = rtl::max(total_sec_max, total_sec);
total_sec_min = rtl::min(total_sec_min, total_sec);
start_sec_max = rtl::max(start_sec_max, start_sec);
start_sec_min = rtl::min(start_sec_min, start_sec);
peak_memory_min = rtl::min(peak_memory_min, instr->memoryinfo.peakOpMemory);
peak_memory_max = rtl::max(peak_memory_max, instr->memoryinfo.peakOpMemory);
width_min = rtl::min(width_min, instr->width);
width_max = rtl::max(width_max, instr->width);
}
}
}
}
if (need_timer) {
AssertEreport((total_sec_max != -1 || start_sec_max != -1) &&
(total_sec_min != HUGE_VAL - 1 || start_sec_min != HUGE_VAL - 1),
MOD_EXECUTOR,
"unexpect min/max value");
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (es->timing)
appendStringInfo(es->str,
" (actual time=[%.3f,%.3f]..[%.3f,%.3f], rows=%.0f)",
start_sec_min,
total_sec_min,
start_sec_max,
total_sec_max,
rows);
else
appendStringInfo(es->str, " (actual rows=%.0f)", rows);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->analyze) {
char str[100] = "\0";
int rc = sprintf_s(str, sizeof(str), "[%.3f,%.3f]", total_sec_min, total_sec_max);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_ROWS, Float8GetDatum(rows));
if (es->timing) {
es->planinfo->m_planInfo->put(ACTUAL_TIME, PointerGetDatum(cstring_to_text(str)));
}
char width[50] = "\0";
if (width_max > 0) {
rc = sprintf_s(width, sizeof(width), "[%d,%d]", width_min, width_max);
securec_check_ss(rc, "\0", "\0");
}
es->planinfo->m_planInfo->put(ACTUAL_WIDTH, PointerGetDatum(cstring_to_text(width)));
if (es->planinfo->m_planInfo->m_query_mem_mode) {
char opmem[50] = "\0";
if (planstate->plan->operatorMemKB[0] > 0) {
if (planstate->plan->operatorMemKB[0] < 1024)
rc = sprintf_s(opmem, sizeof(opmem), "%dKB", (int)planstate->plan->operatorMemKB[0]);
else {
if (planstate->plan->operatorMemKB[0] > MIN_OP_MEM &&
planstate->plan->operatorMaxMem > planstate->plan->operatorMemKB[0])
rc = sprintf_s(opmem,
sizeof(opmem),
"%dMB(%dMB)",
(int)planstate->plan->operatorMemKB[0] / 1024,
(int)planstate->plan->operatorMaxMem / 1024);
else
rc = sprintf_s(
opmem, sizeof(opmem), "%dMB", (int)planstate->plan->operatorMemKB[0] / 1024);
}
securec_check_ss(rc, "\0", "\0");
}
es->planinfo->m_planInfo->put(ESTIMATE_MEMORY, PointerGetDatum(cstring_to_text(opmem)));
}
char memoryPeakMinSize[100] = "\0";
if (peak_memory_min < 1024)
rc = sprintf_s(memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldBYTE", peak_memory_min);
else if (peak_memory_min < 1024 * 1024)
rc = sprintf_s(memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldKB", peak_memory_min / 1024);
else
rc = sprintf_s(
memoryPeakMinSize, sizeof(memoryPeakMinSize), "%ldMB", peak_memory_min / (1024 * 1024));
securec_check_ss(rc, "\0", "\0");
char memoryPeakMaxSize[100] = "\0";
if (peak_memory_max < 1024)
rc = sprintf_s(memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldBYTE", peak_memory_max);
else if (peak_memory_max < 1024 * 1024)
rc = sprintf_s(memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldKB", peak_memory_max / 1024);
else
rc = sprintf_s(
memoryPeakMaxSize, sizeof(memoryPeakMaxSize), "%ldMB", peak_memory_max / (1024 * 1024));
securec_check_ss(rc, "\0", "\0");
char memoryStr[256] = "\0";
rc = sprintf_s(memoryStr, sizeof(memoryStr), "[%s,%s]", memoryPeakMinSize, memoryPeakMaxSize);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_MEMORY, PointerGetDatum(cstring_to_text(memoryStr)));
}
} else {
if (es->timing) {
ExplainPropertyFloat("Actual Min Startup Time", start_sec_min, 3, es);
ExplainPropertyFloat("Actual Max Startup Time", start_sec_max, 3, es);
ExplainPropertyFloat("Actual Min Total Time", total_sec_min, 3, es);
ExplainPropertyFloat("Actual Max Total Time", total_sec_max, 3, es);
}
ExplainPropertyFloat("Actual Total Rows", rows, 0, es);
}
} else if (es->analyze) {
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
char str[100] = "\0";
int rc = sprintf_s(str, sizeof(str), "[0,0]");
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_ROWS, Float8GetDatum(rows));
if (es->timing) {
es->planinfo->m_planInfo->put(ACTUAL_TIME, PointerGetDatum(cstring_to_text(str)));
}
char memoryStr[100] = "\0";
rc = sprintf_s(memoryStr, sizeof(memoryStr), "[0, 0]");
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_planInfo->put(ACTUAL_MEMORY, PointerGetDatum(cstring_to_text(memoryStr)));
} else
appendStringInfo(es->str, " (Actual time: never executed)");
} else {
ExplainPropertyFloat("Actual Startup Time", 0.0, 3, es);
ExplainPropertyFloat("Actual Total Time", 0.0, 3, es);
}
}
if (es->format == EXPLAIN_FORMAT_TEXT)
appendStringInfoChar(es->str, '\n');
}
return;
}
static void show_tidbitmap_info(BitmapHeapScanState *planstate, ExplainState *es)
{
if (es->format != EXPLAIN_FORMAT_TEXT)
{
ExplainPropertyLong("Exact Heap Blocks", planstate->exact_pages, es);
ExplainPropertyLong("Lossy Heap Blocks", planstate->lossy_pages, es);
} else {
if (planstate->exact_pages > 0 || planstate->lossy_pages > 0) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, "Heap Blocks:");
if (planstate->exact_pages > 0) {
appendStringInfo(es->str, " exact=%ld", planstate->exact_pages);
}
if (planstate->lossy_pages > 0) {
appendStringInfo(es->str, " lossy=%ld", planstate->lossy_pages);
}
appendStringInfoChar(es->str, '\n');
}
}
}
* If it's EXPLAIN ANALYZE, show instrumentation information for a plan node
*
* "which" identifies which instrumentation counter to print
*/
static void show_instrumentation_count(const char* qlabel, int which, const PlanState* planstate, ExplainState* es)
{
double nfiltered = 0.0;
Instrumentation* instr = NULL;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (!es->analyze || !planstate->instrument)
return;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
for (int j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
if (which == 1)
nfiltered += instr->nfiltered1;
else if (which == 2)
nfiltered += instr->nfiltered2;
}
}
}
} else {
if (which == 1)
nfiltered = planstate->instrument->nfiltered1;
else if (which == 2)
nfiltered = planstate->instrument->nfiltered2;
}
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL &&
(nfiltered > 0 || es->format != EXPLAIN_FORMAT_TEXT)) {
ExplainPropertyFloat(qlabel, nfiltered, 0, es);
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_detailInfo) {
if (nfiltered > 0 && es->format == EXPLAIN_FORMAT_TEXT) {
char buf[512];
int rc = 0;
rc = snprintf_s(buf, sizeof(buf), sizeof(buf) - 1, "%.*f", 0, nfiltered);
securec_check_ss(rc, "\0", "\0");
es->planinfo->m_detailInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str, "%s: %s\n", qlabel, buf);
}
}
}
* Show removed rows by filters.
*/
static void show_removed_rows(int which, const PlanState* planstate, int idx, int smpIdx, int* removeRows)
{
Instrumentation* instr = NULL;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(idx, planstate->plan->plan_node_id, smpIdx);
if (instr != NULL && instr->nloops > 0) {
if (which == 1) {
*removeRows = check_integer_overflow(instr->nfiltered1);
} else if (which == 2) {
*removeRows = check_integer_overflow(instr->nfiltered2);
}
}
} else {
if (which == 1) {
*removeRows = check_integer_overflow(planstate->instrument->nfiltered1);
} else if (which == 2) {
*removeRows = check_integer_overflow(planstate->instrument->nfiltered2);
}
}
}
* Check for possible integer overflow when assign a double variable to an int variable.
*/
static int check_integer_overflow(double var)
{
if (var > (double)PG_INT32_MAX || var < (double)PG_INT32_MIN) {
ereport(ERROR,
(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE), errmodule(MOD_OPT),
errmsg("Integer overflow."),
errdetail("Integer overflow occurred when assigning a double variable to an int variable."),
errcause("Try to assign a double variable to an int variable."),
erraction("Please check whether the double variable exceeds the representation range of int.")));
}
return (int)var;
}
* Show extra information for a ForeignScan node.
*/
static void show_foreignscan_info(ForeignScanState* fsstate, ExplainState* es)
{
FdwRoutine* fdwroutine = fsstate->fdwroutine;
if (u_sess->attr.attr_common.show_fdw_remote_plan) {
es->es_frs.node_num++;
ExplainPropertyInteger("Node ID", es->es_frs.node_num , es);
}
if (NULL != fdwroutine && NULL != fdwroutine->ExplainForeignScan)
fdwroutine->ExplainForeignScan(fsstate, es);
if (u_sess->attr.attr_common.show_fdw_remote_plan) {
if (es->es_frs.node_num == 1) {
es->es_frs.str=makeStringInfo();
appendStringInfo(es->es_frs.str, "\nFDW remote plans:\n");
}
appendStringInfo(es->es_frs.str, "Node %d: ", es->es_frs.node_num);
if (NULL != fdwroutine && NULL != fdwroutine->ExplainForeignScanRemote) {
fdwroutine->ExplainForeignScanRemote(fsstate, es);
} else {
appendStringInfo(es->es_frs.str, "No remote plan information.\n");
}
}
}
* @Description: check whether Instrumentation object has
* bloomfilter info to display about storage.
* @IN instr: Instrumentation object
* @Return: true if needed to display bloomfilter info; otherwise false.
* @See also:
*/
static inline bool storage_has_bloomfilter_info(const Instrumentation* instr)
{
return (instr->bloomFilterRows > 0);
}
* @Description: check whether Instrumentation object has
* min/max info to display about storage.
* @IN instr: Instrumentation object
* @Return: true if needed to display minmax filter info; otherwise false.
* @See also:
*/
static inline bool storage_has_minmax_filter(const Instrumentation* instr)
{
return (instr->minmaxFilterFiles > 0 || instr->minmaxFilterStripe > 0 || instr->minmaxFilterStride > 0 ||
instr->minmaxFilterRows > 0);
}
* @Description: check whether Instrumentation object has
* pruned-partition info to display about storage.
* @IN instr: Instrumentation object
* @Return: true if needed to display pruned-partition info; otherwise false.
* @See also:
*/
static inline bool storage_has_pruned_info(const Instrumentation* instr)
{
return (instr->dynamicPrunFiles > 0 || instr->staticPruneFiles > 0);
}
* @Description: print filter info in detail in storage level
* @in instr - instrument info
* @in instr_info - the string to be append
* @return - void
*/
static void show_detail_storage_info_json_of_dfs(Instrumentation* instr, StringInfo instr_info, ExplainState* es)
{
if (storage_has_bloomfilter_info(instr)) {
ExplainPropertyLong("rows skipped by bloom filter", instr->bloomFilterRows, es);
if (instr->bloomFilterBlocks > 0) {
ExplainPropertyLong("include strides) ", instr->bloomFilterBlocks, es);
}
}
if (storage_has_minmax_filter(instr)) {
if (instr->minmaxFilterRows > 0) {
ExplainPropertyLong("min max skip rows", instr->minmaxFilterRows, es);
}
if (instr->minmaxFilterStride > 0) {
ExplainPropertyLong("min max filter strides", instr->minmaxFilterStride, es);
ExplainPropertyLong("min max check strides", instr->minmaxCheckStride, es);
}
if (instr->minmaxFilterStripe > 0) {
ExplainPropertyLong("min max filter stripes", instr->minmaxFilterStripe, es);
ExplainPropertyLong("min max check stripes", instr->minmaxCheckStripe, es);
}
if (instr->minmaxFilterFiles > 0) {
ExplainPropertyLong("min max filter files", instr->minmaxFilterFiles, es);
ExplainPropertyLong("min max check files", instr->minmaxCheckFiles, es);
}
}
if (storage_has_pruned_info(instr)) {
if (instr->staticPruneFiles > 0) {
ExplainPropertyLong("pruned files static", instr->staticPruneFiles, es);
}
if (instr->dynamicPrunFiles > 0) {
ExplainPropertyLong("pruned files dynamic", instr->dynamicPrunFiles, es);
}
}
}
static void show_detail_storage_info_json_of_logft(const Instrumentation* instr, StringInfo instr_info, ExplainState* es)
{
if (instr->minmaxFilterStripe > 0) {
ExplainPropertyText("refuted by hostname", "true", es);
}
if (instr->minmaxCheckStride > 0) {
ExplainPropertyLong("dirname total", instr->minmaxCheckStride, es);
if (instr->minmaxFilterStride > 0) {
ExplainPropertyLong("dirname refuted", instr->minmaxFilterStride, es);
ExplainPropertyLong("dirname scan", (instr->minmaxCheckStride - instr->minmaxFilterStride), es);
}
}
if (instr->minmaxCheckFiles > 0) {
ExplainPropertyLong("filename total", instr->minmaxCheckFiles, es);
if (instr->minmaxFilterFiles > 0) {
ExplainPropertyLong("filename refuted", instr->minmaxFilterFiles, es);
ExplainPropertyLong("filename scan", (instr->minmaxCheckFiles - instr->minmaxFilterFiles), es);
}
if (instr->staticPruneFiles > 0) {
ExplainPropertyLong("filename incompleted", instr->staticPruneFiles, es);
}
if (instr->dynamicPrunFiles) {
ExplainPropertyLong("filename latest files", instr->dynamicPrunFiles, es);
}
}
}
static void show_detail_storage_info_json(Instrumentation* instr, StringInfo instr_info, ExplainState* es)
{
if (instr->dfsType == TYPE_LOG_FT) {
show_detail_storage_info_json_of_logft(instr, instr_info, es);
} else {
show_detail_storage_info_json_of_dfs(instr, instr_info, es);
}
}
static void show_detail_storage_info_text_of_dfs(Instrumentation* instr, StringInfo instr_info)
{
if (storage_has_bloomfilter_info(instr)) {
appendStringInfo(instr_info, "(skip %lu rows by bloom filter", instr->bloomFilterRows);
if (instr->bloomFilterBlocks > 0) {
appendStringInfo(instr_info, " include %lu strides) ", instr->bloomFilterBlocks);
} else {
appendStringInfo(instr_info, ") ");
}
}
if (storage_has_minmax_filter(instr)) {
appendStringInfo(instr_info, "(min max skip:");
if (instr->minmaxFilterRows > 0) {
appendStringInfo(instr_info, " rows %lu", instr->minmaxFilterRows);
}
if (instr->minmaxFilterStride > 0) {
appendStringInfo(instr_info, " strides %lu/%lu", instr->minmaxFilterStride, instr->minmaxCheckStride);
}
if (instr->minmaxFilterStripe > 0) {
appendStringInfo(instr_info, " stripes %lu/%lu", instr->minmaxFilterStripe, instr->minmaxCheckStripe);
}
if (instr->minmaxFilterFiles > 0) {
appendStringInfo(instr_info, " files %lu/%lu", instr->minmaxFilterFiles, instr->minmaxCheckFiles);
}
appendStringInfo(instr_info, ")");
}
if (storage_has_pruned_info(instr)) {
appendStringInfo(instr_info, "(pruned files:");
if (instr->staticPruneFiles > 0) {
appendStringInfo(instr_info, " static %lu", instr->staticPruneFiles);
}
if (instr->dynamicPrunFiles > 0) {
appendStringInfo(instr_info, " dynamic %lu", instr->dynamicPrunFiles);
}
appendStringInfo(instr_info, ")");
}
appendStringInfo(instr_info, "\n");
}
static void show_detail_storage_info_text_of_logft(const Instrumentation* instr, StringInfo instr_info)
{
if (instr->minmaxFilterStripe > 0) {
appendStringInfo(instr_info, "(refuted by hostname)");
}
if (instr->minmaxCheckStride > 0) {
appendStringInfo(instr_info, "(dirname: total %lu", instr->minmaxCheckStride);
if (instr->minmaxFilterStride > 0) {
appendStringInfo(instr_info,
", refuted %lu, scan %lu",
instr->minmaxFilterStride,
(instr->minmaxCheckStride - instr->minmaxFilterStride));
}
appendStringInfo(instr_info, ")");
}
if (instr->minmaxCheckFiles > 0) {
appendStringInfo(instr_info, "(filename: total %lu", instr->minmaxCheckFiles);
if (instr->minmaxFilterFiles > 0) {
appendStringInfo(instr_info,
", refuted %lu, scan %lu",
instr->minmaxFilterFiles,
(instr->minmaxCheckFiles - instr->minmaxFilterFiles));
}
if (instr->staticPruneFiles > 0) {
appendStringInfo(instr_info, ", incompleted %lu", instr->staticPruneFiles);
}
if (instr->dynamicPrunFiles) {
appendStringInfo(instr_info, ", latest files %lu", instr->dynamicPrunFiles);
}
appendStringInfo(instr_info, ")");
}
appendStringInfo(instr_info, "\n");
}
static void show_detail_storage_info_text(Instrumentation* instr, StringInfo instr_info)
{
if (instr->dfsType == TYPE_LOG_FT) {
show_detail_storage_info_text_of_logft(instr, instr_info);
} else {
show_detail_storage_info_text_of_dfs(instr, instr_info);
}
}
* @Description: print storage info in detail=off
* @in planstate - current plan state info
* @in es - explain state info
* @return - void
*/
static void show_analyze_storage_info_of_dfs(const PlanState* planstate, ExplainState* es)
{
int i = 0;
int j = 0;
uint64 total_bloomfilter_rows = 0;
uint64 total_dynamicfiles = 0;
uint64 total_staticFiles = 0;
Instrumentation* instr = NULL;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
for (i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL) {
total_bloomfilter_rows += instr->bloomFilterRows;
total_dynamicfiles += instr->dynamicPrunFiles;
total_staticFiles += instr->staticPruneFiles;
}
}
}
if (total_bloomfilter_rows > 0 || total_dynamicfiles > 0 || total_staticFiles > 0) {
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "(skip %lu rows by bloom filter,", total_bloomfilter_rows);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
" pruned files: static %lu, dynamic %lu)\n",
total_dynamicfiles,
total_staticFiles);
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "(skip %lu rows by bloom filter,", total_bloomfilter_rows);
appendStringInfo(
es->str, " pruned files: static %lu, dynamic %lu)\n", total_dynamicfiles, total_staticFiles);
} else {
ExplainPropertyLong("rows skipped by bloom filter", total_bloomfilter_rows, es);
ExplainPropertyLong("pruned files static", total_dynamicfiles, es);
ExplainPropertyLong("pruned files dynamic", total_staticFiles, es);
}
}
}
static void show_analyze_storage_info_of_logft(const PlanState* planstate, ExplainState* es)
{
uint64 total_hostname = 0;
uint64 total_refuted_by_hostname = 0;
uint64 total_dirname = 0;
uint64 total_refuted_by_dirname = 0;
uint64 total_filename = 0;
uint64 total_refuted_by_filename = 0;
uint64 total_incompleted = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
Instrumentation* instr = NULL;
int node_idx = 0;
int dop_idx = 0;
for (node_idx = 0; node_idx < u_sess->instr_cxt.global_instr->getInstruNodeNum(); node_idx++) {
for (dop_idx = 0; dop_idx < dop; dop_idx++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(node_idx, planstate->plan->plan_node_id, dop_idx);
if (instr != NULL) {
total_hostname += instr->minmaxCheckStripe;
total_refuted_by_hostname += instr->minmaxFilterStripe;
total_dirname += instr->minmaxCheckStride;
total_refuted_by_dirname += instr->minmaxFilterStride;
total_filename += instr->minmaxCheckFiles;
total_refuted_by_filename += instr->minmaxFilterFiles;
total_incompleted += instr->staticPruneFiles;
}
}
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_staticInfo) {
es->planinfo->m_staticInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"(nodes refuted %lu, total %lu)",
total_refuted_by_hostname,
total_hostname);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"(dirname refuted %lu, total %lu)",
total_refuted_by_dirname,
total_dirname);
appendStringInfo(es->planinfo->m_staticInfo->info_str,
"(filename refuted %lu, total %lu, incompleted %lu)\n",
total_refuted_by_filename,
total_filename,
total_incompleted);
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "(nodes refuted %lu, total %lu)", total_refuted_by_hostname, total_hostname);
appendStringInfo(es->str, "(dirname refuted %lu, total %lu)", total_refuted_by_dirname, total_dirname);
appendStringInfo(es->str,
"(filename refuted %lu, total %lu, incompleted %lu)\n",
total_refuted_by_filename,
total_filename,
total_incompleted);
} else {
ExplainPropertyLong("nodes refuted", total_refuted_by_hostname, es);
ExplainPropertyLong("nodes total", total_hostname, es);
ExplainPropertyLong("dirname refuted", total_refuted_by_dirname, es);
ExplainPropertyLong("dirname total", total_dirname, es);
ExplainPropertyLong("filename refuted", total_refuted_by_filename, es);
ExplainPropertyLong("filename total", total_filename, es);
ExplainPropertyLong("filename incompleted", total_incompleted, es);
}
}
static void show_analyze_storage_info(PlanState* planstate, ExplainState* es)
{
if (u_sess->instr_cxt.global_instr->getInstruNodeNum() > 0) {
Instrumentation* instr = u_sess->instr_cxt.global_instr->getInstrSlot(0, planstate->plan->plan_node_id, 0);
if (instr && instr->dfsType == TYPE_LOG_FT) {
show_analyze_storage_info_of_logft(planstate, es);
} else {
show_analyze_storage_info_of_dfs(planstate, es);
}
}
}
static void show_storage_filter_info(PlanState* planstate, ExplainState* es)
{
AssertEreport(planstate != NULL && es != NULL, MOD_EXECUTOR, "unexpect null value");
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
int i = 0;
int j = 0;
bool has_info = false;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (es->detail) {
for (i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr == NULL)
continue;
if (storage_has_bloomfilter_info(instr) || storage_has_minmax_filter(instr) ||
storage_has_pruned_info(instr)) {
if (has_info == false)
ExplainOpenGroup("Storage Detail", "Storage Detail", false, es);
has_info = true;
ExplainOpenGroup("Plan", NULL, true, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(i, 0, BLOOM_FILTER_INFO, true);
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
es->planinfo->m_datanodeInfo->set_datanode_name(node_name, j, dop);
show_detail_storage_info_text(instr, es->planinfo->m_datanodeInfo->info_str);
continue;
}
append_datanode_name(es, node_name, dop, j);
if (es->format == EXPLAIN_FORMAT_TEXT)
show_detail_storage_info_text(instr, es->str);
else
show_detail_storage_info_json(instr, es->str, es);
ExplainCloseGroup("Plan", NULL, true, es);
}
}
}
if (has_info)
ExplainCloseGroup("Storage Detail", "Storage Detail", false, es);
} else
show_analyze_storage_info(planstate, es);
}
}
static void show_wlm_explain_name(ExplainState* es, const char* plan_name, const char* pname, int plan_node_id)
{
if (plan_name != NULL) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
appendStringInfo(es->str, "%s\n", plan_name);
es->indent++;
}
int digit = get_digit(plan_node_id);
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit - digit);
appendStringInfo(es->str, "%d | ", plan_node_id);
if (es->indent) {
appendStringInfoSpaces(es->str, es->indent);
appendStringInfoString(es->str, "-> ");
}
appendStringInfoString(es->str, pname);
es->indent++;
}
* Show extra information of MERGE for a ModifyTable node
*/
static void show_modifytable_merge_info(const PlanState* planstate, ExplainState* es)
{
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
Instrumentation* instr = NULL;
if (es->detail) {
ExplainOpenGroup("Merge Detail", "Merge Detail", false, es);
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ExplainOpenGroup("Plan", NULL, true, es);
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id);
if (instr == NULL)
continue;
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
append_datanode_name(es, node_name, 1, 0);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(i, 0, MERGE_INSERTED, instr->nfiltered1);
es->planinfo->m_runtimeinfo->put(i, 0, MERGE_UPDATED, instr->nfiltered2);
es->planinfo->m_datanodeInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_datanodeInfo->info_str, "%s ", node_name);
appendStringInfo(
es->planinfo->m_datanodeInfo->info_str, "(Merge: inserted %.0f, ", instr->nfiltered1);
appendStringInfo(es->planinfo->m_datanodeInfo->info_str, "updated %.0f)\n", instr->nfiltered2);
continue;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, "(Merge: inserted %.0f, ", instr->nfiltered1);
appendStringInfo(es->str, "updated %.0f)\n", instr->nfiltered2);
} else {
ExplainPropertyFloat("Merge Inserted", instr->nfiltered1, 0, es);
ExplainPropertyFloat("Merge Updated", instr->nfiltered2, 0, es);
}
ExplainCloseGroup("Plan", NULL, true, es);
}
ExplainCloseGroup("Merge Detail", "Merge Detail", false, es);
}
}
}
static void ShowRoughCheckInfo(ExplainState* es, struct Instrumentation* instrument, int nodeIdx, int smpIdx)
{
RCInfo* rcPtr = &instrument->rcInfo;
if (instrument->needRCInfo) {
if (es->planinfo && es->planinfo->m_runtimeinfo) {
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, CU_NONE, Int64GetDatum(rcPtr->m_CUNone));
es->planinfo->m_runtimeinfo->put(nodeIdx, smpIdx, CU_SOME, Int64GetDatum(rcPtr->m_CUSome));
appendStringInfoString(es->planinfo->m_datanodeInfo->info_str, "(RoughCheck CU: ");
appendStringInfo(
es->planinfo->m_datanodeInfo->info_str, "CUNone: %lu, CUSome: %lu", rcPtr->m_CUNone, rcPtr->m_CUSome);
appendStringInfoString(es->planinfo->m_datanodeInfo->info_str, ")\n");
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoString(es->str, " (RoughCheck CU: ");
appendStringInfo(es->str, "CUNone: %lu, CUSome: %lu", rcPtr->m_CUNone, rcPtr->m_CUSome);
appendStringInfoString(es->str, ")\n");
} else {
ExplainPropertyLong("RoughCheck CUNone", rcPtr->m_CUNone, es);
ExplainPropertyLong("RoughCheck CUSome", rcPtr->m_CUSome, es);
}
}
}
* Fetch the name of an index in an EXPLAIN
*
* We allow plugins to get control here so that plans involving hypothetical
* indexes can be explained.
*/
static const char* explain_get_index_name(Oid indexId)
{
const char* result = NULL;
if (explain_get_index_name_hook) {
result = (*explain_get_index_name_hook) (indexId);
if(result)
return result;
}
result = get_rel_name(indexId);
if (result == NULL)
ereport(ERROR, (errcode(ERRCODE_CACHE_LOOKUP_FAILED), errmsg("cache lookup failed for index %u", indexId)));
result = quote_identifier(result);
return result;
}
* Add some additional details about an IndexScan, IndexOnlyScan or AnnIndexScan
*/
static void ExplainIndexScanDetails(Oid indexid, ScanDirection indexorderdir, ExplainState* es)
{
const char* indexname = explain_get_index_name(indexid);
if (es->format == EXPLAIN_FORMAT_TEXT) {
if (ScanDirectionIsBackward(indexorderdir))
appendStringInfoString(es->str, " Backward");
appendStringInfo(es->str, " using %s", indexname);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL) {
StringInfo tmpName = &es->planinfo->m_planInfo->m_pname;
if (ScanDirectionIsBackward(indexorderdir))
appendStringInfoString(tmpName, " Backward");
appendStringInfo(tmpName, " using %s", indexname);
}
} else {
const char* scandir = NULL;
switch (indexorderdir) {
case BackwardScanDirection:
scandir = "Backward";
break;
case NoMovementScanDirection:
scandir = "NoMovement";
break;
case ForwardScanDirection:
scandir = "Forward";
break;
default:
scandir = "?\?\?";
break;
}
ExplainPropertyText("Scan Direction", scandir, es);
ExplainPropertyText("Index Name", indexname, es);
}
}
* Show the target of a Scan node
*/
static void ExplainScanTarget(Scan* plan, ExplainState* es)
{
ExplainTargetRel((Plan*)plan, plan->scanrelid, es);
}
* Show the target of a ModifyTable node
*/
static void ExplainModifyTarget(ModifyTable* plan, ExplainState* es)
{
Index rti;
bool multiTarget = (list_length((List*)linitial(plan->resultRelations)) > 1);
* We show the name of the first target relation. In multi-target-table
* cases this should always be the parent of the inheritance tree.
*/
AssertEreport(plan->resultRelations != NIL, MOD_EXECUTOR, "unexpect empty list");
rti = (Index)linitial_int((List*)linitial(plan->resultRelations));
ExplainTargetRel((Plan*)plan, rti, es, multiTarget);
}
* Show the target relation of a scan or modify node
*/
static void ExplainTargetRel(Plan* plan, Index rti, ExplainState* es, bool multiTarget)
{
char* objectname = NULL;
char* namespc = NULL;
const char* objecttag = NULL;
char* object_type = NULL;
RangeTblEntry* rte = NULL;
if (IsA(plan, RemoteQuery) || IsA(plan, VecRemoteQuery)) {
RemoteQuery* rq = (RemoteQuery*)plan;
if (rq->is_simple)
return;
}
rte = rt_fetch(rti, es->rtable);
if (rte == NULL) {
ereport(ERROR,
(errcode(ERRCODE_UNEXPECTED_NULL_VALUE),
errmsg("Unable to get relation.")));
}
if (es->plan && rte) {
es->planinfo->m_planTableData->set_object_type(rte, &object_type);
}
switch (nodeTag(plan)) {
case T_ModifyTable:
case T_VecModifyTable: {
if (multiTarget) {
objecttag = "Relation Name";
break;
}
Assert(rte != NULL);
Assert(rte->rtekind == RTE_RELATION);
objectname = get_rel_name(rte->relid);
if (es->verbose || es->plan)
namespc = get_namespace_name(get_rel_namespace(rte->relid));
objecttag = "Relation Name";
} break;
case T_SeqScan:
case T_CStoreScan:
#ifdef ENABLE_HTAP
case T_IMCStoreScan:
#endif
#ifdef ENABLE_MULTIPLE_NODES
case T_TsStoreScan:
#endif
case T_IndexScan:
case T_IndexOnlyScan:
case T_AnnIndexScan:
case T_BitmapHeapScan:
case T_CStoreIndexScan:
case T_CStoreIndexCtidScan:
case T_CStoreIndexHeapScan:
case T_TidScan:
case T_TidRangeScan:
case T_ForeignScan:
case T_ExtensiblePlan:
case T_VecForeignScan: {
Assert(rte != NULL);
Assert(rte->rtekind == RTE_RELATION);
objectname = get_rel_name(rte->relid);
if (es->verbose || es->plan)
namespc = get_namespace_name(get_rel_namespace(rte->relid));
objecttag = "Relation Name";
} break;
case T_FunctionScan: {
Node* funcexpr = NULL;
Assert(rte != NULL && rte->rtekind == RTE_FUNCTION);
* If the expression is still a function call, we can get the
* real name of the function. Otherwise, punt (this can
* happen if the optimizer simplified away the function call,
* for example).
*/
funcexpr = ((FunctionScan*)plan)->funcexpr;
if (funcexpr && IsA(funcexpr, FuncExpr)) {
Oid funcid = ((FuncExpr*)funcexpr)->funcid;
objectname = get_func_name(funcid);
if (es->verbose || es->plan)
namespc = get_namespace_name(get_func_namespace(funcid));
}
objecttag = "Function Name";
} break;
case T_ValuesScan: {
Assert(rte != NULL);
Assert(rte->rtekind == RTE_VALUES);
} break;
case T_CteScan: {
Assert(rte != NULL && rte->rtekind == RTE_CTE && !rte->self_reference);
objectname = rte->ctename;
objecttag = "CTE Name";
} break;
case T_WorkTableScan: {
Assert(rte != NULL && rte->rtekind == RTE_CTE && rte->self_reference);
objectname = rte->ctename;
objecttag = "CTE Name";
} break;
case T_RemoteQuery: {
RemoteQuery* rq = (RemoteQuery*)plan;
if (rq->is_simple == false && rte) {
Assert(rte->rtekind == RTE_REMOTE_DUMMY);
objectname = rte->relname;
objecttag = "RemoteQuery name";
}
} break;
default:
break;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
StringInfo tmpName;
if (t_thrd.explain_cxt.explain_perf_mode == EXPLAIN_NORMAL)
tmpName = es->str;
else
tmpName = &es->planinfo->m_planInfo->m_pname;
appendStringInfoString(tmpName, " on");
if (multiTarget) {
appendStringInfoString(tmpName, " MULTI-RELATION");
} else {
if (namespc != NULL && objectname != NULL) {
appendStringInfo(tmpName, " %s.%s", quote_identifier(namespc), quote_identifier(objectname));
} else if (objectname != NULL) {
appendStringInfo(tmpName, " %s", quote_identifier(objectname));
}
if (rte && rte->eref && (objectname == NULL || strcmp(rte->eref->aliasname, objectname) != 0)) {
appendStringInfo(tmpName, " %s", quote_identifier(rte->eref->aliasname));
}
if (IsA(plan, CStoreScan) && ((CStoreScan*)plan)->minMaxInfo != NULL) {
appendStringInfo(tmpName, " %s", "(min-max optimization)");
}
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL)
es->planinfo->m_planInfo->put(PLAN, PointerGetDatum(cstring_to_text(tmpName->data)));
} else {
if (objecttag != NULL && multiTarget) {
ExplainPropertyText(objecttag, "MULTI-RELATION", es);
} else {
if (objecttag != NULL && objectname != NULL)
ExplainPropertyText(objecttag, objectname, es);
if (namespc != NULL)
ExplainPropertyText("Schema", namespc, es);
if (rte != NULL && rte->eref != NULL) {
ExplainPropertyText("Alias", rte->eref->aliasname, es);
}
}
}
if (es->plan) {
if (multiTarget) {
es->planinfo->m_planTableData->set_plan_table_objs(plan->plan_node_id, "MULTI-RELATION", NULL, NULL);
return;
}
if (objectname == NULL && rte != NULL && rte->eref != NULL)
objectname = rte->eref->aliasname;
if (objectname != NULL)
es->planinfo->m_planTableData->set_plan_table_objs(plan->plan_node_id, objectname, object_type, namespc);
}
}
* Show extra information for upsert info
*/
static void show_on_duplicate_info(ModifyTableState* mtstate, ExplainState* es, List* ancestors)
{
ResultRelInfo* resultRelInfo = mtstate->resultRelInfo;
IndexInfo* indexInfo = NULL;
List* idxNames = NIL;
for (int i = 0; i < resultRelInfo->ri_NumIndices; ++i) {
indexInfo = resultRelInfo->ri_IndexRelationInfo[i];
if (!indexInfo->ii_Unique && !indexInfo->ii_ExclusionOps) {
continue;
}
Relation indexRelation = resultRelInfo->ri_IndexRelationDescs[i];
char* indexName = RelationGetRelationName(indexRelation);
idxNames = lappend(idxNames, indexName);
}
ExplainPropertyText("Conflict Resolution",
mtstate->mt_upsert->us_action == UPSERT_NOTHING ? "NOTHING" : "UPDATE",
es);
* Don't display arbiter indexes at all when DO NOTHING variant
* implicitly ignores all conflicts
*/
if (idxNames != NIL) {
ExplainPropertyList("Conflict Arbiter Indexes", idxNames, es);
}
if (mtstate->mt_upsert->us_updateWhere != NIL) {
ModifyTable *node = (ModifyTable*)mtstate->ps.plan;
Node* expr = NULL;
if (IsA(node->upsertWhere, List)) {
expr = (Node*)make_ands_explicit((List*)node->upsertWhere);
} else {
expr = node->upsertWhere;
}
List* clauseList = list_make1(expr);
show_upper_qual(clauseList, "Conflict Filter", &mtstate->ps, ancestors, es);
list_free(clauseList);
show_instrumentation_count("Rows Removed by Conflict Filter", 1, &mtstate->ps, es);
}
}
#ifndef PGXC
* Show extra information for a ModifyTable node
*/
static void show_modifytable_info(ModifyTableState* mtstate, ExplainState* es)
{
FdwRoutine* fdwroutine = mtstate->resultRelInfo->ri_FdwRoutine;
* If the first target relation is a foreign table, call its FDW to
* display whatever additional fields it wants to. For now, we ignore the
* possibility of other targets being foreign tables, although the API for
* ExplainForeignModify is designed to allow them to be processed.
*/
if (fdwroutine != NULL && fdwroutine->ExplainForeignModify != NULL) {
ModifyTable* node = (ModifyTable*)mtstate->ps.plan;
List* fdw_private = (List*)linitial(node->fdwPrivLists);
fdwroutine->ExplainForeignModify(mtstate, mtstate->resultRelInfo, fdw_private, 0, es);
}
}
#endif
* Explain the constituent plans of a ModifyTable, Append, MergeAppend,
* BitmapAnd, or BitmapOr node.
*
* The ancestors list should already contain the immediate parent of these
* plans.
*
* Note: we don't actually need to examine the Plan list members, but
* we need the list in order to determine the length of the PlanState array.
*/
static void ExplainMemberNodes(const List* plans, PlanState** planstates, List* ancestors, ExplainState* es)
{
int nplans = list_length(plans);
int j;
bool old_flag = es->from_dn;
for (j = 0; j < nplans; j++) {
es->from_dn = old_flag;
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL)
ExplainNode<true>(planstates[j], ancestors, "Member", NULL, es);
else
ExplainNode<false>(planstates[j], ancestors, "Member", NULL, es);
}
es->from_dn = old_flag;
}
* Explain a list of SubPlans (or initPlans, which also use SubPlan nodes).
*
* The ancestors list should already contain the immediate parent of these
* SubPlanStates.
*/
static void ExplainSubPlans(List* plans, List* ancestors, const char* relationship, ExplainState* es)
{
ListCell* lst = NULL;
bool old_flag = es->from_dn;
foreach (lst, plans) {
SubPlanState* sps = (SubPlanState*)lfirst(lst);
SubPlan* sp = (SubPlan*)sps->xprstate.expr;
es->from_dn = old_flag;
if (sps->planstate == NULL)
continue;
if (sps->planstate->plan->exec_type == EXEC_ON_DATANODES)
es->from_dn = true;
* For stream planning of RecursiveUnion, we already print the RecursiveUnion
* plan node in normal position, so in subplan list, we skip this part
*
* e.g.
*
* id | operation
* ----+---------------------------------------------------------------------------------
* 1 | -> Streaming (type: GATHER)
* 2 | -> Sort
* 3 | -> CTE Scan on rq
* 4 | -> Recursive Union
* 5 | -> Seq Scan on chinamap
* 6 | -> Hash Join (7,9)
* 7 | -> Streaming(type: REDISTRIBUTE) stream_level:1
* 8 | -> Seq Scan on chinamap origin <<ruid:[4] ctlid:[4] (SYNC)>>
* 9 | -> Hash
* 10 | -> Streaming(type: REDISTRIBUTE) stream_level:1
* 11 | -> WorkTable Scan on rq <<ruid:[4] ctlid:[4]>>
*/
if (STREAM_RECURSIVECTE_SUPPORTED && sp->subLinkType == CTE_SUBLINK) {
continue;
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo != NULL)
ExplainNode<true>(sps->planstate, ancestors, relationship, sp->plan_name, es);
else
ExplainNode<false>(sps->planstate, ancestors, relationship, sp->plan_name, es);
}
es->from_dn = old_flag;
}
static void ExplainPrettyList(List* data, ExplainState* es)
{
ListCell* lc = NULL;
bool first = true;
foreach (lc, data) {
if (!first) {
appendStringInfoString(es->planinfo->m_verboseInfo->info_str, ", ");
}
appendStringInfoString(es->planinfo->m_verboseInfo->info_str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->planinfo->m_verboseInfo->info_str, '\n');
}
* Explain a property like what ExplainPropertyList does but append it to
* the end of the plan tree.
* We reuse ExplainPropertyList here by temporarily setting and recovering
* es->str such that the explation is printed into es->post_str instead.
* "data" is a list of C strings.
*/
void ExplainPropertyListPostPlanTree(const char* qlabel, List* data, ExplainState* es)
{
if (es->post_str == NULL) {
es->post_str = makeStringInfo();
}
StringInfo str_backup = es->str;
int indent_backup = es->indent;
es->str = es->post_str;
es->indent = 0;
ExplainPropertyList(qlabel, data, es);
es->str = str_backup;
es->indent = indent_backup;
}
* Explain a property, such as sort keys or targets, that takes the form of
* a list of unlabeled items. "data" is a list of C strings.
*/
void ExplainPropertyList(const char* qlabel, List* data, ExplainState* es)
{
ListCell* lc = NULL;
bool first = true;
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
}
appendStringInfo(es->str, "%s: ", qlabel);
foreach (lc, data) {
if (!first)
appendStringInfoString(es->str, ", ");
appendStringInfoString(es->str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->str, '\n');
break;
case EXPLAIN_FORMAT_XML:
ExplainXMLTag(qlabel, X_OPENING, es);
foreach (lc, data) {
char* str = NULL;
appendStringInfoSpaces(es->str, es->indent * 2 + 2);
appendStringInfoString(es->str, "<Item>");
str = escape_xml((const char*)lfirst(lc));
appendStringInfoString(es->str, str);
pfree_ext(str);
appendStringInfoString(es->str, "</Item>\n");
}
ExplainXMLTag(qlabel, X_CLOSING, es);
break;
case EXPLAIN_FORMAT_JSON:
ExplainJSONLineEnding(es);
appendStringInfoSpaces(es->str, es->indent * 2);
escape_json(es->str, qlabel);
appendStringInfoString(es->str, ": [");
foreach (lc, data) {
if (!first)
appendStringInfoString(es->str, ", ");
escape_json(es->str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->str, ']');
break;
case EXPLAIN_FORMAT_YAML:
ExplainYAMLLineStarting(es);
appendStringInfo(es->str, "%s: ", qlabel);
foreach (lc, data) {
appendStringInfoChar(es->str, '\n');
appendStringInfoSpaces(es->str, es->indent * 2 + 2);
appendStringInfoString(es->str, "- ");
escape_yaml(es->str, (const char*)lfirst(lc));
}
break;
default:
break;
}
}
* Explain a property that takes the form of a list of unlabeled items within
* another list. "data" is a list of C strings.
*/
void ExplainPropertyListNested(const char* qlabel, List* data, ExplainState* es)
{
ListCell* lc = NULL;
bool first = true;
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
case EXPLAIN_FORMAT_XML:
ExplainPropertyList(qlabel, data, es);
return;
case EXPLAIN_FORMAT_JSON:
ExplainJSONLineEnding(es);
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfoChar(es->str, '[');
foreach (lc, data) {
if (!first)
appendStringInfoString(es->str, ", ");
escape_json(es->str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->str, ']');
break;
case EXPLAIN_FORMAT_YAML:
ExplainYAMLLineStarting(es);
appendStringInfoString(es->str, "- [");
foreach (lc, data) {
if (!first)
appendStringInfoString(es->str, ", ");
escape_yaml(es->str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->str, ']');
break;
default:
break;
}
}
* Explain a simple property.
*
* If "numeric" is true, the value is a number (or other value that
* doesn't need quoting in JSON).
*
* This usually should not be invoked directly, but via one of the datatype
* specific routines ExplainPropertyText, ExplainPropertyInteger, etc.
*/
static void ExplainProperty(const char* qlabel, const char* value, bool numeric, ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
if (es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, *es->wlm_statistics_plan_max_digit);
appendStringInfoString(es->str, " | ");
appendStringInfoSpaces(es->str, es->indent);
} else {
appendStringInfoSpaces(es->str, es->indent * 2);
}
appendStringInfo(es->str, "%s: %s\n", qlabel, value);
break;
case EXPLAIN_FORMAT_XML: {
char* str = NULL;
appendStringInfoSpaces(es->str, es->indent * 2);
ExplainXMLTag(qlabel, X_OPENING | X_NOWHITESPACE, es);
str = escape_xml(value);
appendStringInfoString(es->str, str);
pfree_ext(str);
ExplainXMLTag(qlabel, X_CLOSING | X_NOWHITESPACE, es);
appendStringInfoChar(es->str, '\n');
} break;
case EXPLAIN_FORMAT_JSON:
ExplainJSONLineEnding(es);
appendStringInfoSpaces(es->str, es->indent * 2);
escape_json(es->str, qlabel);
appendStringInfoString(es->str, ": ");
if (numeric)
appendStringInfoString(es->str, value);
else
escape_json(es->str, value);
break;
case EXPLAIN_FORMAT_YAML:
ExplainYAMLLineStarting(es);
appendStringInfo(es->str, "%s: ", qlabel);
if (numeric)
appendStringInfoString(es->str, value);
else
escape_yaml(es->str, value);
break;
default:
break;
}
}
* Explain a string-valued property.
*/
void ExplainPropertyText(const char* qlabel, const char* value, ExplainState* es)
{
ExplainProperty(qlabel, value, false, es);
}
* Explain an integer-valued property.
*/
void ExplainPropertyInteger(const char* qlabel, int value, ExplainState* es)
{
char buf[32];
errno_t rc = snprintf_s(buf, sizeof(buf), sizeof(buf) - 1, "%d", value);
securec_check_ss(rc, "\0", "\0");
ExplainProperty(qlabel, buf, true, es);
}
* Explain a long-integer-valued property.
*/
void ExplainPropertyLong(const char* qlabel, long value, ExplainState* es)
{
char buf[32];
errno_t rc = snprintf_s(buf, sizeof(buf), sizeof(buf) - 1, "%ld", value);
securec_check_ss(rc, "\0", "\0");
ExplainProperty(qlabel, buf, true, es);
}
* Explain a float-valued property, using the specified number of
* fractional digits.
*/
void ExplainPropertyFloat(const char* qlabel, double value, int ndigits, ExplainState* es)
{
char buf[256];
errno_t rc = snprintf_s(buf, sizeof(buf), sizeof(buf) - 1, "%.*f", ndigits, value);
securec_check_ss(rc, "\0", "\0");
ExplainProperty(qlabel, buf, true, es);
}
* Open a group of related objects.
*
* objtype is the type of the group object, labelname is its label within
* a containing object (if any).
*
* If labeled is true, the group members will be labeled properties,
* while if it's false, they'll be unlabeled objects.
*/
static void ExplainOpenGroup(const char* objtype, const char* labelname, bool labeled, ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
break;
case EXPLAIN_FORMAT_XML:
ExplainXMLTag(objtype, X_OPENING, es);
es->indent++;
break;
case EXPLAIN_FORMAT_JSON:
ExplainJSONLineEnding(es);
appendStringInfoSpaces(es->str, 2 * es->indent);
if (labelname != NULL) {
escape_json(es->str, labelname);
appendStringInfoString(es->str, ": ");
}
appendStringInfoChar(es->str, labeled ? '{' : '[');
* In JSON format, the grouping_stack is an integer list. 0 means
* we've emitted nothing at this grouping level, 1 means we've
* emitted something (and so the next item needs a comma). See
* ExplainJSONLineEnding().
*/
es->grouping_stack = lcons_int(0, es->grouping_stack);
es->indent++;
break;
case EXPLAIN_FORMAT_YAML:
* In YAML format, the grouping stack is an integer list. 0 means
* we've emitted nothing at this grouping level AND this grouping
* level is unlabelled and must be marked with "- ". See
* ExplainYAMLLineStarting().
*/
ExplainYAMLLineStarting(es);
if (labelname != NULL) {
appendStringInfo(es->str, "%s: ", labelname);
es->grouping_stack = lcons_int(1, es->grouping_stack);
} else {
appendStringInfoString(es->str, "- ");
es->grouping_stack = lcons_int(0, es->grouping_stack);
}
es->indent++;
break;
default:
break;
}
}
* Close a group of related objects.
* Parameters must match the corresponding ExplainOpenGroup call.
*/
static void ExplainCloseGroup(const char* objtype, const char* labelname, bool labeled, ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
break;
case EXPLAIN_FORMAT_XML:
es->indent--;
ExplainXMLTag(objtype, X_CLOSING, es);
break;
case EXPLAIN_FORMAT_JSON:
es->indent--;
appendStringInfoChar(es->str, '\n');
appendStringInfoSpaces(es->str, 2 * es->indent);
appendStringInfoChar(es->str, labeled ? '}' : ']');
es->grouping_stack = list_delete_first(es->grouping_stack);
break;
case EXPLAIN_FORMAT_YAML:
es->indent--;
es->grouping_stack = list_delete_first(es->grouping_stack);
break;
default:
break;
}
}
* Emit a "dummy" group that never has any members.
*
* objtype is the type of the group object, labelname is its label within
* a containing object (if any).
*/
static void ExplainDummyGroup(const char* objtype, const char* labelname, ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
break;
case EXPLAIN_FORMAT_XML:
ExplainXMLTag(objtype, X_CLOSE_IMMEDIATE, es);
break;
case EXPLAIN_FORMAT_JSON:
ExplainJSONLineEnding(es);
appendStringInfoSpaces(es->str, 2 * es->indent);
if (labelname != NULL) {
escape_json(es->str, labelname);
appendStringInfoString(es->str, ": ");
}
escape_json(es->str, objtype);
break;
case EXPLAIN_FORMAT_YAML:
ExplainYAMLLineStarting(es);
if (labelname != NULL) {
escape_yaml(es->str, labelname);
appendStringInfoString(es->str, ": ");
} else {
appendStringInfoString(es->str, "- ");
}
escape_yaml(es->str, objtype);
break;
default:
break;
}
}
* Emit the start-of-output boilerplate.
*
* This is just enough different from processing a subgroup that we need
* a separate pair of subroutines.
*/
void ExplainBeginOutput(ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
break;
case EXPLAIN_FORMAT_XML:
appendStringInfoString(es->str, "<explain xmlns=\"http://www.postgresql.org/2009/explain\">\n");
es->indent++;
break;
case EXPLAIN_FORMAT_JSON:
appendStringInfoChar(es->str, '[');
es->grouping_stack = lcons_int(0, es->grouping_stack);
es->indent++;
break;
case EXPLAIN_FORMAT_YAML:
es->grouping_stack = lcons_int(0, es->grouping_stack);
break;
default:
break;
}
}
* Emit the end-of-output boilerplate.
*/
void ExplainEndOutput(ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
break;
case EXPLAIN_FORMAT_XML:
es->indent--;
appendStringInfoString(es->str, "</explain>");
break;
case EXPLAIN_FORMAT_JSON:
es->indent--;
appendStringInfoString(es->str, "\n]");
es->grouping_stack = list_delete_first(es->grouping_stack);
break;
case EXPLAIN_FORMAT_YAML:
es->grouping_stack = list_delete_first(es->grouping_stack);
break;
default:
break;
}
}
* Put an appropriate separator between multiple plans
*/
void ExplainSeparatePlans(ExplainState* es)
{
switch (es->format) {
case EXPLAIN_FORMAT_TEXT:
appendStringInfoChar(es->str, '\n');
break;
case EXPLAIN_FORMAT_XML:
case EXPLAIN_FORMAT_JSON:
case EXPLAIN_FORMAT_YAML:
break;
default:
break;
}
}
#ifdef PGXC
* Emit execution node list number.
*/
static void ExplainExecNodes(const ExecNodes* en, ExplainState* es)
{
int primary_node_count = en ? list_length(en->primarynodelist) : 0;
int node_count = en ? list_length(en->nodeList) : 0;
if (!es->num_nodes)
return;
if (es->format == EXPLAIN_FORMAT_TEXT) {
appendStringInfo(es->str, " (primary node count=%d, node count=%d)", primary_node_count, node_count);
} else {
ExplainPropertyInteger("Primary node count", primary_node_count, es);
ExplainPropertyInteger("Node count", node_count, es);
}
}
static void showStreamnetwork(Stream* stream, ExplainState* es)
{
Instrumentation* instr = NULL;
Plan* plan = (Plan*)stream;
long spacePeakKb = 0;
double toltal_time = 0;
double deserialize_time = 0;
double copy_time = 0;
if (u_sess->instr_cxt.global_instr && es->planinfo->m_runtimeinfo) {
int dop = ((Plan*)stream)->parallel_enabled ? u_sess->opt_cxt.query_dop : 1;
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
#ifdef ENABLE_MULTIPLE_NODES
char* node_name = PGXCNodeGetNodeNameFromId(i, PGXC_NODE_DATANODE);
#else
char* node_name = NULL;
if (!IS_SPQ_RUNNING) {
node_name = g_instance.exec_cxt.nodeName;
} else {
node_name = (char*)palloc0(SPQNODENAMELEN);
sprintf(node_name, "%d", i);
}
#endif
for (int j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, plan->plan_node_id, j);
if (instr == NULL)
continue;
spacePeakKb = (instr->network_perfdata.network_size + 1023) / 1024;
toltal_time = instr->network_perfdata.total_poll_time;
deserialize_time = instr->network_perfdata.total_deserialize_time;
copy_time = instr->network_perfdata.total_copy_time;
es->planinfo->m_runtimeinfo->put(i, j, QUERY_NETWORK, spacePeakKb);
es->planinfo->m_runtimeinfo->put(i, j, NETWORK_POLL_TIME, Float8GetDatum(toltal_time));
es->planinfo->m_query_summary->m_size += instr->network_perfdata.network_size;
es->planinfo->m_staticInfo->set_plan_name<false, true>();
es->planinfo->m_staticInfo->set_datanode_name(node_name, j, dop);
if (!STREAM_IS_LOCAL_NODE(stream->smpDesc.distriType))
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Stream Network: %ldkB, ", spacePeakKb);
appendStringInfo(es->planinfo->m_staticInfo->info_str, "Network Poll Time: %.3f", toltal_time);
if (copy_time > 0.0)
appendStringInfo(es->planinfo->m_staticInfo->info_str, "; Data Copy Time: %.3f\n", copy_time);
else
appendStringInfo(
es->planinfo->m_staticInfo->info_str, "; Data Deserialize Time: %.3f\n", deserialize_time);
}
}
}
}
* ShowRunNodeInfo
* show exec node information, including group name and node list
*
* @param (in) en:
* the exec nodes
* @param (in) es:
* the explain state
* @param (in) qlabel:
* the leading label of this item in explain information
*
* @return: void
*/
static void ShowRunNodeInfo(const ExecNodes* en, ExplainState* es, const char* qlabel)
{
if (en != NULL && es->nodes) {
StringInfo node_names = makeStringInfo();
#ifdef ENABLE_MULTIPLE_NODES
if (list_length(en->nodeList) == u_sess->pgxc_cxt.NumDataNodes) {
appendStringInfo(node_names, "All datanodes");
} else {
ListCell* lcell = NULL;
char* sep = "";
int node_no;
if (ng_enable_nodegroup_explain()) {
char* group_name = ng_get_dist_group_name((Distribution*)&(en->distribution));
appendStringInfo(node_names, "(%s) ", group_name);
}
foreach (lcell, en->nodeList) {
NameData nodename = {{0}};
node_no = lfirst_int(lcell);
appendStringInfo(node_names,
"%s%s",
sep,
get_pgxc_nodename(PGXCNodeGetNodeOid(node_no, PGXC_NODE_DATANODE), &nodename));
sep = ", ";
}
}
#else
appendStringInfo(node_names, "All datanodes");
#endif
ExplainPropertyText(qlabel, node_names->data, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "%s: %s\n", qlabel, node_names->data);
}
}
}
* ShowStreamRunNodeInfo
* show the stream thread spawn on which node,
* also include consumer nodes.
*
* @param (in) stream:
* the stream node
* @param (in) es:
* the explain state
*
* @return: void
*/
static void ShowStreamRunNodeInfo(Stream* stream, ExplainState* es)
{
ExecNodes* en = stream->scan.plan.exec_nodes;
ShowRunNodeInfo(en, es, "Spawn on");
if (es->nodes && es->verbose) {
ExecNodes* cen = stream->consumer_nodes;
ShowRunNodeInfo(cen, es, "Consumer Nodes");
}
}
* Emit remote query planning details
*/
static void ExplainRemoteQuery(RemoteQuery* plan, PlanState* planstate, List* ancestors, ExplainState* es)
{
ExecNodes* en = plan->exec_nodes;
const char* label_name = NULL;
if (es->format == EXPLAIN_FORMAT_TEXT) {
label_name = "Node/s";
} else {
label_name = "Nodes";
}
if (en != NULL && es->nodes && plan->position == GATHER) {
StringInfo node_names = makeStringInfo();
ListCell* lcell = NULL;
char* sep = NULL;
int node_no;
if (en->primarynodelist) {
if (plan->exec_type == EXEC_ON_ALL_NODES) {
appendStringInfo(node_names, "All nodes");
} else {
char exec_node_type =
(plan->exec_type == EXEC_ON_DATANODES) ? (PGXC_NODE_DATANODE) : (PGXC_NODE_COORDINATOR);
sep = "";
foreach (lcell, en->primarynodelist) {
NameData nodename = {{0}};
node_no = lfirst_int(lcell);
appendStringInfo(node_names,
"%s%s",
sep,
get_pgxc_nodename(PGXCNodeGetNodeOid(node_no, exec_node_type), &nodename));
sep = ", ";
}
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
ExplainPropertyText("Primary node/s", node_names->data, es);
} else {
ExplainPropertyText("Primary nodes", node_names->data, es);
}
}
if (en->nodeList) {
resetStringInfo(node_names);
if (list_length(en->nodeList) == u_sess->pgxc_cxt.NumDataNodes) {
appendStringInfo(node_names, "All datanodes");
es->datanodeinfo.all_datanodes = true;
} else {
char exec_node_type =
(plan->exec_type == EXEC_ON_DATANODES) ? (PGXC_NODE_DATANODE) : (PGXC_NODE_COORDINATOR);
es->datanodeinfo.all_datanodes = false;
es->datanodeinfo.len_nodelist = list_length(en->nodeList);
if (es->datanodeinfo.node_index) {
pfree_ext(es->datanodeinfo.node_index);
es->datanodeinfo.node_index = NULL;
}
if (es->datanodeinfo.node_index == NULL)
es->datanodeinfo.node_index = (int*)palloc0(sizeof(int) * es->datanodeinfo.len_nodelist);
sep = "";
if (ng_enable_nodegroup_explain()) {
char* group_name = ng_get_dist_group_name(&(en->distribution));
appendStringInfo(node_names, "(%s) ", group_name);
}
int i = 0;
foreach (lcell, en->nodeList) {
NameData nodename = {{0}};
node_no = lfirst_int(lcell);
es->datanodeinfo.node_index[i++] = node_no;
appendStringInfo(node_names,
"%s%s",
sep,
get_pgxc_nodename(PGXCNodeGetNodeOid(node_no, exec_node_type), &nodename));
sep = ", ";
}
}
ExplainPropertyText(label_name, node_names->data, es);
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "Node/s: %s\n", node_names->data);
}
}
}
if (plan->position == SCAN_GATHER) {
ExplainPropertyText(label_name, "DNs in compute pool", es);
}
if (plan->position == PLAN_ROUTER) {
StringInfo address = makeStringInfo();
Plan* fsplan = get_foreign_scan((Plan*)plan);
ServerTypeOption srvtype = getServerType(((ForeignScan*)fsplan)->scan_relid);
if (T_OBS_SERVER == srvtype || T_TXT_CSV_OBS_SERVER == srvtype) {
ComputePoolConfig** conf = get_cp_conninfo();
appendStringInfo(address, "%s:%s", conf[0]->cpip, conf[0]->cpport);
} else {
DummyServerOptions* options = getDummyServerOption();
AssertEreport(options != NULL, MOD_EXECUTOR, "unexpect null value");
appendStringInfo(address, "%s", options->address);
AssertEreport(address != NULL, MOD_EXECUTOR, "unexpect null value");
}
StringInfo si = makeStringInfo();
appendStringInfo(si, "compute pool: %s", address->data);
ExplainPropertyText(label_name, si->data, es);
}
if (en != NULL && en->en_expr && es->pstmt->commandType != CMD_MERGE)
show_expression((Node*)en->en_expr, "Node expr", planstate, ancestors, es->verbose, es);
else if (en != NULL && en->en_expr && es->pstmt->commandType == CMD_MERGE) {
ListCell* lc = NULL;
int cnt = 0;
StringInfo si = makeStringInfo();
foreach (lc, en->en_expr) {
if (IsA(lfirst(lc), Var)) {
Var* var = (Var*)lfirst(lc);
if (cnt == 0)
appendStringInfo(si, "$%d", var->varattno);
else
appendStringInfo(si, ", $%d", var->varattno);
cnt++;
}
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_detailInfo) {
es->planinfo->m_detailInfo->set_plan_name<true, true>();
appendStringInfo(es->planinfo->m_detailInfo->info_str, "%s: %s\n", "Node expr: ", si->data);
}
ExplainPropertyText("Node expr", si->data, es);
}
if (es->verbose && !plan->is_simple)
ExplainPropertyText("Remote query", plan->sql_statement, es);
}
#endif
* Emit opening or closing XML tag.
*
* "flags" must contain X_OPENING, X_CLOSING, or X_CLOSE_IMMEDIATE.
* Optionally, OR in X_NOWHITESPACE to suppress the whitespace we'd normally
* add.
*
* XML tag names can't contain white space, so we replace any spaces in
* "tagname" with dashes.
*/
static void ExplainXMLTag(const char* tagname, unsigned int flags, ExplainState* es)
{
const char* s = NULL;
if ((flags & X_NOWHITESPACE) == 0)
appendStringInfoSpaces(es->str, 2 * es->indent);
appendStringInfoCharMacro(es->str, '<');
if ((flags & X_CLOSING) != 0)
appendStringInfoCharMacro(es->str, '/');
for (s = tagname; *s; s++)
appendStringInfoCharMacro(es->str, (*s == ' ') ? '-' : *s);
if ((flags & X_CLOSE_IMMEDIATE) != 0)
appendStringInfoString(es->str, " /");
appendStringInfoCharMacro(es->str, '>');
if ((flags & X_NOWHITESPACE) == 0)
appendStringInfoCharMacro(es->str, '\n');
}
* Emit a JSON line ending.
*
* JSON requires a comma after each property but the last. To facilitate this,
* in JSON format, the text emitted for each property begins just prior to the
* preceding line-break (and comma, if applicable).
*/
static void ExplainJSONLineEnding(ExplainState* es)
{
AssertEreport(es->format == EXPLAIN_FORMAT_JSON, MOD_EXECUTOR, "unexpect explain state format");
if (linitial_int(es->grouping_stack) != 0)
appendStringInfoChar(es->str, ',');
else
linitial_int(es->grouping_stack) = 1;
appendStringInfoChar(es->str, '\n');
}
* Indent a YAML line.
*
* YAML lines are ordinarily indented by two spaces per indentation level.
* The text emitted for each property begins just prior to the preceding
* line-break, except for the first property in an unlabelled group, for which
* it begins immediately after the "- " that introduces the group. The first
* property of the group appears on the same line as the opening "- ".
*/
static void ExplainYAMLLineStarting(ExplainState* es)
{
AssertEreport(es->format == EXPLAIN_FORMAT_YAML, MOD_EXECUTOR, "unexpect explain state format");
if (linitial_int(es->grouping_stack) == 0) {
linitial_int(es->grouping_stack) = 1;
} else {
appendStringInfoChar(es->str, '\n');
appendStringInfoSpaces(es->str, es->indent * 2);
}
}
* YAML is a superset of JSON; unfortuantely, the YAML quoting rules are
* ridiculously complicated -- as documented in sections 5.3 and 7.3.3 of
* http://yaml.org/spec/1.2/spec.html -- so we chose to just quote everything.
* Empty strings, strings with leading or trailing whitespace, and strings
* containing a variety of special characters must certainly be quoted or the
* output is invalid; and other seemingly harmless strings like "0xa" or
* "true" must be quoted, lest they be interpreted as a hexadecimal or Boolean
* constant rather than a string.
*/
static void escape_yaml(StringInfo buf, const char* str)
{
escape_json(buf, str);
}
#ifndef ENABLE_MULTIPLE_NODES
static void SetPredictor(ExplainState* es, PlanTable* m_planInfo)
{
if (es->opt_model_name != NULL) {
char *labels = NULL;
int nLabel;
(void) CheckModelTargets(RLSTM_TEMPLATE_NAME, es->opt_model_name, &labels, &nLabel);
for (int i = 0; i < nLabel; ++i) {
switch (labels[i]) {
case gs_opt_model_label_total_time: {
m_planInfo->m_pred_time = true;
break;
}
case gs_opt_model_label_rows: {
m_planInfo->m_pred_row = true;
break;
}
case gs_opt_model_label_peak_memory: {
m_planInfo->m_pred_mem = true;
break;
}
default:
break;
}
}
}
}
#endif
void PlanInformation::init(ExplainState* es, int plansize, int num_nodes, bool query_mem_mode)
{
m_detail = es->detail;
m_planInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_planInfo->m_costs = es->costs;
m_planInfo->m_verbose = es->verbose;
m_planInfo->m_cpu = false;
m_planInfo->m_analyze = es->analyze;
m_planInfo->m_timing = es->timing;
m_planInfo->m_pred_time = false;
m_planInfo->m_pred_row = false;
m_planInfo->m_pred_mem = false;
#ifndef ENABLE_MULTIPLE_NODES
SetPredictor(es, m_planInfo);
#endif
if (m_planInfo->m_costs)
m_planInfo->m_query_mem_mode = query_mem_mode;
m_planInfo->init(PLANINFO);
m_planInfo->init_planinfo(plansize);
m_detailInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_detailInfo->init(DETAILINFO);
m_query_summary = (PlanTable*)palloc0(sizeof(PlanTable));
m_query_summary->init(QUERYSUMMARY);
if (es->plan) {
m_planTableData = (PlanTable*)palloc0(sizeof(PlanTable));
if (IsExplainPlanSelectForUpdateStmt && plansize == 0)
plansize = 1;
m_planTableData->init_plan_table_data(plansize);
} else
m_planTableData = NULL;
if (es->analyze) {
m_staticInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_staticInfo->init(STATICINFO);
m_profileInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_profileInfo->init(PROFILEINFO);
} else {
m_staticInfo = NULL;
m_profileInfo = NULL;
}
if (es->verbose) {
m_verboseInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_verboseInfo->init(VERBOSEINFO);
} else
m_verboseInfo = NULL;
if (es->detail) {
m_datanodeInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_datanodeInfo->init(DATANODEINFO);
m_runtimeinfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_runtimeinfo->init(RUNTIMEINFO);
m_runtimeinfo->init_multi_info(es, plansize, num_nodes);
} else {
m_datanodeInfo = NULL;
m_runtimeinfo = NULL;
}
if ((es->cpu || es->buffers) && es->detail == false) {
m_planInfo->m_cpu = true;
m_IOInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_IOInfo->init(IOINFO);
} else
m_IOInfo = NULL;
if (es->detail) {
m_recursiveInfo = (PlanTable*)palloc0(sizeof(PlanTable));
m_recursiveInfo->init(RECURSIVEINFO);
} else {
m_recursiveInfo = NULL;
}
}
void PlanInformation::append_str_info(const char* data, int id, const char* value)
{
if (m_verboseInfo)
appendStringInfo(m_verboseInfo->info_str, data, id, value);
if (m_datanodeInfo)
appendStringInfo(m_datanodeInfo->info_str, data, id, value);
}
static void writefile(const char* path, char* text)
{
FILE* out_file = NULL;
char* line = NULL;
errno_t rc = 0;
if ((out_file = fopen(path, "w")) == NULL) {
return;
}
while (*text) {
char* eol = NULL;
int len;
eol = strchr(text, '\n');
if (eol != NULL) {
len = eol - text;
eol++;
} else {
len = strlen(text);
eol += len;
}
line = (char*)palloc0(sizeof(char) * (len + 2));
rc = memcpy_s(line, len, text, len);
securec_check(rc, "\0", "\0");
line[len] = '\n';
if (fputs(line, out_file) < 0) {
(void)fclose(out_file);
return;
}
pfree_ext(line);
text = eol;
}
if (fclose(out_file)) {
return;
}
}
void PlanInformation::dump_runtimeinfo_file()
{
int rc = 0;
char tempdirpath[1024] = {0};
FILE* fp = NULL;
bool is_absolute = false;
if (u_sess->attr.attr_common.Perf_log == NULL) {
is_absolute = is_absolute_path(u_sess->attr.attr_common.Log_directory);
if (is_absolute)
rc = snprintf_s(tempdirpath,
sizeof(tempdirpath),
sizeof(tempdirpath) - 1,
"%s/explain_table_%d.csv",
u_sess->attr.attr_common.Log_directory,
m_query_id);
else
rc = snprintf_s(tempdirpath,
sizeof(tempdirpath),
sizeof(tempdirpath) - 1,
"%s/pg_log/explain_table_%d.csv",
g_instance.attr.attr_common.data_directory,
m_query_id);
} else {
rc = snprintf_s(
tempdirpath, sizeof(tempdirpath), sizeof(tempdirpath) - 1, "%s", u_sess->attr.attr_common.Perf_log);
}
securec_check_ss(rc, "\0", "\0");
fp = fopen(tempdirpath, PG_BINARY_W);
if (fp == NULL)
return;
fclose(fp);
m_runtimeinfo->flush_data_to_file();
writefile(tempdirpath, m_runtimeinfo->info_str->data);
}
void PlanInformation::free_memory()
{
if (m_planInfo) {
pfree_ext(m_planInfo);
m_planInfo = NULL;
}
if (m_detailInfo) {
pfree_ext(m_detailInfo);
m_detailInfo = NULL;
}
if (m_staticInfo) {
pfree_ext(m_staticInfo);
m_staticInfo = NULL;
}
if (m_verboseInfo) {
pfree_ext(m_verboseInfo);
m_verboseInfo = NULL;
}
if (m_datanodeInfo) {
pfree_ext(m_datanodeInfo);
m_datanodeInfo = NULL;
}
if (m_IOInfo) {
pfree_ext(m_IOInfo);
m_IOInfo = NULL;
}
if (m_profileInfo) {
pfree_ext(m_profileInfo);
m_profileInfo = NULL;
}
if (m_query_summary) {
pfree_ext(m_query_summary);
m_query_summary = NULL;
}
if (m_recursiveInfo) {
pfree(m_recursiveInfo);
m_recursiveInfo = NULL;
}
}
int PlanInformation::print_plan(Portal portal, DestReceiver* dest)
{
m_count = 0;
if (m_planTableData) {
free_memory();
return m_count;
}
if (m_planInfo)
m_count += m_planInfo->print_plan(portal, dest);
if (m_detailInfo && m_detailInfo->m_size > 0)
m_count += m_detailInfo->print_plan(portal, dest);
if (m_staticInfo && m_staticInfo->m_size > 0)
m_count += m_staticInfo->print_plan(portal, dest);
if (m_verboseInfo && m_verboseInfo->m_size > 0)
m_count += m_verboseInfo->print_plan(portal, dest);
if (m_datanodeInfo && m_datanodeInfo->m_size > 0)
m_count += m_datanodeInfo->print_plan(portal, dest);
if (m_IOInfo && m_IOInfo->m_size)
m_count += m_IOInfo->print_plan(portal, dest);
if (m_recursiveInfo && m_recursiveInfo->m_size)
m_count += m_recursiveInfo->print_plan(portal, dest);
if (m_profileInfo && m_profileInfo->m_size > 0)
m_count += m_profileInfo->print_plan(portal, dest);
if (m_query_summary && m_query_summary->m_size > 0)
m_count += m_query_summary->print_plan(portal, dest);
free_memory();
return m_count;
}
void PlanTable::init(int plantype)
{
switch (plantype) {
case PLANINFO:
m_desc = getTupleDesc();
initStringInfo(&m_pname);
break;
case DETAILINFO:
m_desc = getTupleDesc("Predicate Information (identified by plan id)");
info_str = makeStringInfo();
break;
case STATICINFO:
m_desc = getTupleDesc("Memory Information (identified by plan id)");
info_str = makeStringInfo();
break;
case VERBOSEINFO:
m_desc = getTupleDesc("Targetlist Information (identified by plan id)");
info_str = makeStringInfo();
break;
case DATANODEINFO:
m_desc = getTupleDesc("Datanode Information (identified by plan id)");
info_str = makeStringInfo();
break;
case IOINFO:
m_desc = getTupleDesc("Cpu-Buffer Information (identified by plan id)");
info_str = makeStringInfo();
break;
case RUNTIMEINFO:
m_desc = getTupleDesc_detail();
info_str = makeStringInfo();
break;
case PROFILEINFO:
m_desc = getTupleDesc("User Define Profiling");
info_str = makeStringInfo();
break;
case RECURSIVEINFO:
m_desc = getTupleDesc("Recursive Union Plan node (identified by plan id)");
info_str = makeStringInfo();
break;
case QUERYSUMMARY:
m_desc = getTupleDesc("====== Query Summary =====");
info_str = makeStringInfo();
break;
default:
ereport(ERROR, (errcode(ERRCODE_UNRECOGNIZED_NODE_TYPE), errmsg("unsupported type %d", plantype)));
}
}
void PlanTable::init_multi_info(ExplainState* es, int plansize, int num_nodes)
{
int i = 0;
int j = 0;
m_plan_size = plansize;
m_data_size = num_nodes;
m_multi_info = (MultiInfo***)palloc0(sizeof(MultiInfo**) * m_data_size);
for (i = 0; i < m_data_size; i++) {
m_multi_info[i] = (MultiInfo**)palloc0(sizeof(MultiInfo*) * m_plan_size);
for (j = 0; j < m_plan_size; j++) {
int dop = u_sess->instr_cxt.global_instr->getNodeDop(j + 1);
m_multi_info[i][j] = (MultiInfo*)palloc0(sizeof(MultiInfo) * dop);
init_multi_info_data(m_multi_info[i][j], dop);
}
}
}
* @Description: init the data of MultiInfo
*
* @param[IN] multi_info: each element of m_multi_info
* @param[IN] dop: dop of specific operator
* @return: void
*/
void PlanTable::init_multi_info_data(MultiInfo* multi_info, int dop)
{
for (int i = 0; i < dop; i++) {
multi_info[i].m_Datum = (Datum*)palloc0(EXPLAIN_TOTAL_ATTNUM * sizeof(Datum));
multi_info[i].m_Nulls = (bool*)palloc0(EXPLAIN_TOTAL_ATTNUM * sizeof(bool));
for (int j = 0; j < EXPLAIN_TOTAL_ATTNUM; j++)
multi_info[i].m_Nulls[j] = true;
}
}
void PlanTable::init_planinfo(int plansize)
{
int i = 0;
int j = 0;
m_size = plansize;
int attnum;
int analyze_num = m_analyze ? 1 : 0;
int costs_num = m_costs ? 1 : 0;
int verbose_num = m_verbose ? 1 : 0;
attnum = (PLAN + 1) + analyze_num * NUM_ANALYZE_COLS + \
costs_num * (NUM_COST_COLS + verbose_num * NUM_VER_COLS + m_pred_time + m_pred_row + m_pred_mem);
if (m_costs && !m_query_mem_mode)
attnum = attnum - 1;
m_data = (Datum**)palloc0(plansize * sizeof(Datum*));
m_isnull = (bool**)palloc0(plansize * sizeof(bool*));
for (i = 0; i < plansize; i++) {
m_data[i] = (Datum*)palloc0(attnum * sizeof(Datum));
m_isnull[i] = (bool*)palloc0(attnum * sizeof(bool));
for (j = 0; j < attnum; j++)
m_isnull[i][j] = true;
}
}
* Description: Init data struct for plan table according to plan size.
*
* Parameters:
* @in num_plan_nodes: number of plan node.
* Return: void
*/
void PlanTable::init_plan_table_data(int num_plan_nodes)
{
int i = 0;
errno_t rc = EOK;
if (num_plan_nodes == 0)
return;
m_plan_node_num = num_plan_nodes;
m_plan_table = (PlanTableMultiData**)palloc0(num_plan_nodes * sizeof(PlanTableMultiData*));
for (i = 0; i < num_plan_nodes; i++) {
m_plan_table[i] = (PlanTableMultiData*)palloc0(sizeof(PlanTableMultiData));
m_plan_table[i]->m_datum = (PlanTableData*)palloc0(sizeof(PlanTableData));
m_plan_table[i]->m_isnull = (bool*)palloc0(PLANTABLECOLNUM * sizeof(bool));
rc = memset_s(m_plan_table[i]->m_isnull, PLANTABLECOLNUM * sizeof(bool), 1, PLANTABLECOLNUM * sizeof(bool));
securec_check(rc, "\0", "\0");
}
}
TupleDesc PlanTable::getTupleDesc()
{
TupleDesc tupdesc;
int attnum;
int cur = 1;
int analyze_num = m_analyze ? 1 : 0;
int costs_num = m_costs ? 1 : 0;
int verbose_num = m_verbose ? 1 : 0;
attnum = (PLAN + 1) + analyze_num * NUM_ANALYZE_COLS + \
costs_num * (NUM_COST_COLS + verbose_num * NUM_VER_COLS + m_pred_time + m_pred_row + m_pred_mem);
if (m_costs && !m_query_mem_mode)
attnum = attnum - 1;
if (!m_timing && m_analyze) {
attnum = attnum - 1;
}
tupdesc = CreateTemplateTupleDesc(attnum, false);
for (int i = 0; i < SLOT_NUMBER; i++) {
bool add_slot = false;
if (i <= PLAN)
add_slot = true;
else if (m_analyze && plan_table_cols[i].disoption == ANAL_OPT) {
if (!m_timing && plan_table_cols[i].val == ACTUAL_TIME) {
add_slot = false;
} else {
add_slot = true;
}
} else if (m_costs && plan_table_cols[i].disoption == COST_OPT) {
if (m_query_mem_mode || i != ESTIMATE_MEMORY)
add_slot = true;
} else if (m_costs && (m_verbose && plan_table_cols[i].disoption == VERBOSE_OPT)) {
add_slot = true;
} else if (m_costs && (m_pred_time && plan_table_cols[i].disoption == PRED_OPT_TIME)) {
add_slot = true;
} else if (m_costs && (m_pred_row && plan_table_cols[i].disoption == PRED_OPT_ROW)) {
add_slot = true;
} else if (m_costs && (m_pred_mem && plan_table_cols[i].disoption == PRED_OPT_MEM)) {
add_slot = true;
}
if (add_slot) {
TupleDescInitEntry(tupdesc, (AttrNumber)cur, plan_table_cols[i].name, plan_table_cols[i].typid, -1, 0);
m_col_loc[i] = cur - 1;
cur++;
} else
m_col_loc[i] = -1;
}
return tupdesc;
}
TupleDesc PlanTable::getTupleDesc_detail()
{
TupleDesc tupdesc;
int attnum = EXPLAIN_TOTAL_ATTNUM;
int i = 1;
tupdesc = CreateTemplateTupleDesc(attnum, false);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "query id", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "plan parent node id", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "plan node id", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "plan type", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "from datanode", BOOLOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "node name", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "plan name", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "start time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "total time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "oper time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "plan rows", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "loops", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "ex cyc", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "inc cyc", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "ex cyc/rows", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "memory", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "peak query memory", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "shared hit", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "shared read", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "shared dirtied", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "shared written", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "local hit", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "local read", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "local dirtied", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "local written", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "temp read", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "temp written", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "blk read time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "blk write time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "CU None", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "CU Some", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "sort method", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "sort type", TEXTOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "sort space", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "hash batch", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "hash batch orignal", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "hash bucket", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "hash space", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "hash file number", INT4OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "query network", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "network poll time", FLOAT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "llvm optimization", BOOLOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "bloom filter", BOOLOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "dfs block info", BOOLOID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "merge inserted", INT8OID, -1, 0);
TupleDescInitEntry(tupdesc, (AttrNumber)i++, "merge updated", INT8OID, -1, 0);
return tupdesc;
}
TupleDesc PlanTable::getTupleDesc(const char* attname)
{
TupleDesc tupdesc;
tupdesc = CreateTemplateTupleDesc(1, false);
TupleDescInitEntry(tupdesc, (AttrNumber)1, attname, TEXTOID, -1, 0);
return tupdesc;
}
void PlanTable::put(int type, Datum value)
{
AssertEreport(type < SLOT_NUMBER && m_col_loc[type] >= 0, MOD_EXECUTOR, "unexpect input value");
m_data[m_plan_node_id - 1][m_col_loc[type]] = value;
m_isnull[m_plan_node_id - 1][m_col_loc[type]] = false;
}
void PlanTable::put(int nodeId, int smpId, int type, Datum value)
{
if (nodeId == -1) {
for (int i = 0; i < m_data_size; i++) {
int dop = u_sess->instr_cxt.global_instr->getNodeDop(m_plan_node_id);
for (int j = 0; j < dop; j++) {
m_multi_info[i][m_plan_node_id - 1][j].m_Datum[type] = value;
m_multi_info[i][m_plan_node_id - 1][j].m_Nulls[type] = false;
}
}
} else {
m_multi_info[nodeId][m_plan_node_id - 1][smpId].m_Datum[type] = value;
m_multi_info[nodeId][m_plan_node_id - 1][smpId].m_Nulls[type] = false;
}
}
void PlanTable::flush(DestReceiver* dest, int plantype)
{
TupOutputState* tstate = NULL;
tstate = begin_tup_output_tupdesc(dest, m_desc);
switch (plantype) {
case PLANINFO:
flush_plan(tstate);
break;
case DETAILINFO:
case STATICINFO:
case VERBOSEINFO:
case DATANODEINFO:
case IOINFO:
case PROFILEINFO:
case RECURSIVEINFO:
case QUERYSUMMARY:
flush_other(tstate);
break;
default:
ereport(ERROR, (errcode(ERRCODE_UNRECOGNIZED_NODE_TYPE), errmsg("unsupported type %d", plantype)));
}
end_tup_output(tstate);
}
int PlanTable::print_plan(Portal portal, DestReceiver* dest)
{
TupleTableSlot* slot = NULL;
int current_tuple_count = 0;
portal->tupDesc = m_desc;
slot = MakeSingleTupleTableSlot(portal->tupDesc);
current_tuple_count = RunGetSlotFromExplain(portal, slot, dest, m_size);
ExecDropSingleTupleTableSlot(slot);
return current_tuple_count;
}
void PlanTable::flush_plan(TupOutputState* tstate)
{
int i = 0;
for (i = 0; i < m_size; i++) {
do_tup_output(tstate,
m_data[i],
tstate->slot->tts_tupleDescriptor->natts,
m_isnull[i],
tstate->slot->tts_tupleDescriptor->natts);
pfree_ext(m_data[i]);
pfree_ext(m_isnull[i]);
}
pfree_ext(m_data);
pfree_ext(m_isnull);
pfree_ext(m_pname.data);
}
void PlanTable::flush_other(TupOutputState* tstate)
{
m_size = do_text_output_multiline(tstate, info_str->data);
pfree_ext(info_str->data);
}
void PlanTable::set_pname(char* data)
{
m_pname.data = data;
}
* @Description: set datanode name
* @in nodename - current datanode name to be set
* @in smp_idx - current smp index
* @in dop - degree of parallel, 1 means not parallel
* @out - void
*/
void PlanTable::set_datanode_name(char* nodename, int smp_idx, int dop)
{
if (dop == 1)
appendStringInfo(info_str, "%s ", nodename);
else
appendStringInfo(info_str, "%s[worker %d] ", nodename, smp_idx);
}
template <bool set_name, bool set_space>
void PlanTable::set_plan_name()
{
if (set_name) {
if (m_has_write_planname == false) {
appendStringInfo(info_str, "%3d --%s\n", m_plan_node_id, m_pname.data);
m_has_write_planname = true;
}
}
if (set_space) {
appendStringInfoSpaces(info_str, 8);
}
}
template void PlanTable::set_plan_name<true, true>();
* Description: Transform the vaules into upper case.
* Parameters:
* @in str: string wait transform.
* @in len: transform length.
* Return: void
*/
static char* set_strtoupper(const char* str, uint32 len)
{
if (str == NULL)
return NULL;
uint32 out_len = 0;
if (len < strlen(str))
out_len = len;
else
out_len = strlen(str);
char* ptrout = (char*)palloc0(out_len + 1);
for (uint32 i = 0; i < out_len; i++)
ptrout[i] = pg_toupper((unsigned char)str[i]);
ptrout[out_len] = '\0';
return ptrout;
}
* Description: Make session id and prepare to insert.
* Session id is make up of thread start time and thread pid.
* Parameters:
* @in sessid: session_id.
* Return: void
*/
void PlanTable::make_session_id(char* sessid)
{
AssertEreport(sessid != NULL, MOD_EXECUTOR, "unexpect null value");
errno_t rc = EOK;
rc = snprintf_s(sessid,
SESSIONIDLEN,
SESSIONIDLEN - 1,
"%ld.%lu",
IS_THREAD_POOL_WORKER ? u_sess->proc_cxt.MyProcPort->SessionStartTime : t_thrd.proc_cxt.MyStartTime,
IS_THREAD_POOL_WORKER ? u_sess->session_id : t_thrd.proc_cxt.MyProcPid);
securec_check_ss(rc, "\0", "\0");
}
* Description: Set session_id, user_id, statement_id, query_id and plan node id for the query.
* Parameters:
* @in query_id: query id of current plan.
* @in es: explain state.
* Return: void
*/
void PlanTable::set_plan_table_ids(uint64 query_id, ExplainState* es)
{
for (int i = 0; i < m_plan_node_num; i++) {
make_session_id(m_plan_table[i]->m_datum->session_id);
m_plan_table[i]->m_isnull[PT_SESSION_ID] = false;
m_plan_table[i]->m_datum->user_id = GetCurrentUserId();
m_plan_table[i]->m_isnull[PT_USER_ID] = false;
if (es->statement_id != NULL) {
uint32 s_len = strlen(es->statement_id);
if (s_len > (STMTIDLEN - 1))
ereport(ERROR,
(errcode(ERRCODE_DATA_EXCEPTION),
errmsg("statement_id is too long. Input statement_id length=%u, however max length=%d.",
s_len,
STMTIDLEN - 1)));
errno_t rc = strncpy_s(m_plan_table[i]->m_datum->statement_id, STMTIDLEN, es->statement_id, s_len);
securec_check(rc, "\0", "\0");
m_plan_table[i]->m_isnull[PT_STATEMENT_ID] = false;
}
m_plan_table[i]->m_datum->query_id = u_sess->debug_query_id;
m_plan_table[i]->m_isnull[PT_QUERY_ID] = false;
m_plan_table[i]->m_datum->node_id = i + 1;
m_plan_table[i]->m_isnull[PT_NODE_ID] = false;
}
}
* Description: Set operation and options for remote query and stream node.
* Parameters:
* @in pname: plain description for node.
* @out operation: node operation.
* @out options: node options.
* Return: void
*/
void PlanTable::set_plan_table_streaming_ops(char* pname, char** operation, char** options)
{
errno_t rc = EOK;
char* start = NULL;
char* end = NULL;
if (strncasecmp("Streaming", pname, strlen("Streaming")) == 0 ||
strncasecmp("Vector Streaming", pname, strlen("Vector Streaming")) == 0) {
start = strrchr(pname, ':') + 2;
end = strrchr(pname, ')');
uint32 option_len = end - start;
*options = (char*)palloc0(option_len + 1);
rc = memcpy_s(*options, option_len + 1, start, option_len);
securec_check(rc, "\0", "\0");
if (pname[0] == 'S')
*operation = "Streaming";
else
*operation = "Vector Streaming";
}
}
* Description: For join node, we may need to set option 'CARTESIAN' to fit with A db.
* Parameters:
* @in plan: plan node.
* @int options: node options.
* Return: void
*/
void PlanTable::set_plan_table_join_options(Plan* plan, char** options)
{
switch (nodeTag(plan)) {
case T_NestLoop:
case T_VecNestLoop: {
Join* nestloop = (Join*)plan;
if (nestloop->joinqual == NULL)
*options = "CARTESIAN";
} break;
case T_VecMergeJoin:
case T_MergeJoin: {
MergeJoin* mergejoin = (MergeJoin*)plan;
if (mergejoin->mergeclauses == NULL)
*options = "CARTESIAN";
} break;
case T_HashJoin:
case T_VecHashJoin: {
HashJoin* hashjoin = (HashJoin*)plan;
if (hashjoin->hashclauses == NULL)
*options = "CARTESIAN";
} break;
case T_AsofJoin:
case T_VecAsofJoin: {
VecAsofJoin* asofjoin = (VecAsofJoin*)plan;
if (asofjoin->hashclauses == NULL || asofjoin->mergeclauses == NULL)
*options = "CARTESIAN";
} break;
default:
break;
}
}
* Description: Set operation and options for node except stream node.
* Parameters:
* @in plan_node_id: plan node id.
* @int operation: node operation.
* @int options: node options.
* Return: void
*/
void PlanTable::set_plan_table_ops(int plan_node_id, char* operation, char* options)
{
errno_t rc = EOK;
* for query shipping we do not collect plan info.
* However we need collect plan info for select for update.
*/
if (plan_node_id == 0) {
if (IsExplainPlanSelectForUpdateStmt)
plan_node_id = 1;
else
return;
}
if (operation != NULL) {
operation = set_strtoupper(operation, OPERATIONLEN);
rc = strncpy_s(m_plan_table[plan_node_id - 1]->m_datum->operation, OPERATIONLEN, operation, OPERATIONLEN - 1);
securec_check(rc, "\0", "\0");
pfree(operation);
m_plan_table[plan_node_id - 1]->m_isnull[PT_OPERATION] = false;
}
if (options != NULL) {
options = set_strtoupper(options, OPTIONSLEN);
rc = strncpy_s(m_plan_table[plan_node_id - 1]->m_datum->options, OPTIONSLEN, options, OPTIONSLEN - 1);
securec_check(rc, "\0", "\0");
pfree(options);
m_plan_table[plan_node_id - 1]->m_isnull[PT_OPTIONS] = false;
}
}
* Description: Set object_type according to relkind in pg_class.
* Parameters:
* @in rte: related ralation.
* @out object_type: object type.
* Return: void
*/
void PlanTable::set_object_type(RangeTblEntry* rte, char** object_type)
{
if (rte == NULL)
return;
if (rte->rtekind == RTE_RELATION) {
switch (rte->relkind) {
case 'r':
*object_type = "TABLE";
break;
case 'i':
*object_type = "INDEX";
break;
case 'S':
*object_type = "SEQUENCE";
break;
case 't':
*object_type = "TOASTVALUE";
break;
case 'v':
*object_type = "VIEW";
break;
case 'c':
*object_type = "COMPOSITE TYPE";
break;
case 'f':
*object_type = "FOREIGN TABLE";
break;
case 'u':
*object_type = "UNCATALOGED";
break;
default:
break;
}
} else {
switch (rte->rtekind) {
case RTE_SUBQUERY:
*object_type = "SUBQUERY";
break;
case RTE_JOIN:
*object_type = "JOIN";
break;
case RTE_FUNCTION:
*object_type = "FUNCTION";
break;
case RTE_VALUES:
*object_type = "VALUES";
break;
case RTE_CTE:
*object_type = "CTE";
break;
#ifdef PGXC
case RTE_REMOTE_DUMMY:
*object_type = "REMOTE_QUERY";
break;
#endif
default:
break;
}
}
}
* Description: Set object_name, object_type, object_owner for one paln node.
* Parameters:
* @in plan_node_id: plan node id.
* @in object_name: object name.
* @in object_type: object type.
* @in object_owner: object schema.
* Return: void
*/
void PlanTable::set_plan_table_objs(
int plan_node_id, const char* object_name, const char* object_type, const char* object_owner)
{
AssertEreport(plan_node_id <= m_plan_node_num, MOD_EXECUTOR, "plan_node_id exceeds max value");
errno_t rc = EOK;
* for query shipping we do not collect plan info.
* However we need collect plan info for select for update.
*/
if (plan_node_id == 0) {
if (IsExplainPlanSelectForUpdateStmt)
plan_node_id = 1;
else
return;
}
if (object_name != NULL) {
rc = memset_s(m_plan_table[plan_node_id - 1]->m_datum->object_name, NAMEDATALEN, '0', NAMEDATALEN);
securec_check(rc, "\0", "\0");
rc = strncpy_s(m_plan_table[plan_node_id - 1]->m_datum->object_name, NAMEDATALEN, object_name, NAMEDATALEN - 1);
securec_check(rc, "\0", "\0");
m_plan_table[plan_node_id - 1]->m_isnull[PT_OBJECT_NAME] = false;
}
if (object_type != NULL) {
rc = memset_s(m_plan_table[plan_node_id - 1]->m_datum->object_type, OBJECTLEN, '0', OBJECTLEN);
securec_check(rc, "\0", "\0");
rc = strncpy_s(m_plan_table[plan_node_id - 1]->m_datum->object_type, OBJECTLEN, object_type, OBJECTLEN - 1);
securec_check(rc, "\0", "\0");
m_plan_table[plan_node_id - 1]->m_isnull[PT_OBJECT_TYPE] = false;
}
if (object_owner != NULL) {
rc = memset_s(m_plan_table[plan_node_id - 1]->m_datum->object_owner, NAMEDATALEN, '0', NAMEDATALEN);
securec_check(rc, "\0", "\0");
rc = strncpy_s(
m_plan_table[plan_node_id - 1]->m_datum->object_owner, NAMEDATALEN, object_owner, NAMEDATALEN - 1);
securec_check(rc, "\0", "\0");
m_plan_table[plan_node_id - 1]->m_isnull[PT_OBJECT_OWNER] = false;
}
}
* Description: Set targetlist into column 'projection' for one plan node.
* Parameters:
* @in plan_node_id: plan node id.
* @in tlist: targetlist of node.
* Return: void
*/
void PlanTable::set_plan_table_projection(int plan_node_id, List* tlist)
{
if (tlist == NULL)
return;
ListCell* lc = NULL;
bool first = true;
* for query shipping we do not collect plan info.
* However we need collect plan info for select for update.
*/
if (plan_node_id == 0) {
if (IsExplainPlanSelectForUpdateStmt)
plan_node_id = 1;
else
return;
}
m_plan_table[plan_node_id - 1]->m_datum->projection = makeStringInfo();
foreach (lc, tlist) {
if (m_plan_table[plan_node_id - 1]->m_datum->projection->len + strlen((const char*)lfirst(lc)) > PROJECTIONLEN)
break;
if (!first)
appendStringInfoString(m_plan_table[plan_node_id - 1]->m_datum->projection, ", ");
appendStringInfoString(m_plan_table[plan_node_id - 1]->m_datum->projection, (const char*)lfirst(lc));
first = false;
}
m_plan_table[plan_node_id - 1]->m_isnull[PT_PROJECTION] = false;
}
* Description: Set cost and cardinality into PlanTable.
* Parameters:
* @in plan_cost: plan cost.
* @in plan_cardinality: rows accessed by current operation.
* Return: void
*/
void PlanTable::set_plan_table_cost_card(int plan_node_id, double plan_cost, double plan_cardinality)
{
m_plan_table[plan_node_id - 1]->m_datum->cost = plan_cost;
m_plan_table[plan_node_id - 1]->m_datum->cardinality = plan_cardinality;
m_plan_table[plan_node_id - 1]->m_isnull[PT_COST] = false;
m_plan_table[plan_node_id - 1]->m_isnull[PT_CARDINALITY] = false;
}
* Description: Call heap_insert to insert all nodes tuples of the plan into table.
* Parameters:
* Return: void
*/
void PlanTable::insert_plan_table_tuple()
{
const char* plan_table_name = T_PLAN_TABLE_DATA;
Relation plan_table;
TupleDesc plan_table_des;
HeapTuple tuple;
Datum new_record[PLANTABLECOLNUM];
errno_t rc = EOK;
Oid plan_table_oid = RelnameGetRelid(plan_table_name);
if (plan_table_oid == InvalidOid)
ereport(ERROR, (errcode(ERRCODE_INVALID_NAME), errmsg("Relation 'plan_table_data' does not exist")));
plan_table = heap_open(plan_table_oid, RowExclusiveLock);
plan_table_des = RelationGetDescr(plan_table);
for (int i = 0; i < m_plan_node_num; i++) {
* Build a tuple to insert
*/
rc = memset_s(new_record, sizeof(new_record), 0, sizeof(new_record));
securec_check(rc, "\0", "\0");
new_record[PT_SESSION_ID] = CStringGetTextDatum(m_plan_table[i]->m_datum->session_id);
new_record[PT_USER_ID] = ObjectIdGetDatum(m_plan_table[i]->m_datum->user_id);
new_record[PT_STATEMENT_ID] = CStringGetTextDatum(m_plan_table[i]->m_datum->statement_id);
new_record[PT_QUERY_ID] = UInt64GetDatum(m_plan_table[i]->m_datum->query_id);
new_record[PT_NODE_ID] = Int32GetDatum(m_plan_table[i]->m_datum->node_id);
new_record[PT_OPERATION] = CStringGetTextDatum(m_plan_table[i]->m_datum->operation);
new_record[PT_OPTIONS] = CStringGetTextDatum(m_plan_table[i]->m_datum->options);
new_record[PT_OBJECT_NAME] = NameGetDatum(m_plan_table[i]->m_datum->object_name);
new_record[PT_OBJECT_TYPE] = CStringGetTextDatum(m_plan_table[i]->m_datum->object_type);
new_record[PT_OBJECT_OWNER] = NameGetDatum(m_plan_table[i]->m_datum->object_owner);
if (m_plan_table[i]->m_datum->projection != NULL)
new_record[PT_PROJECTION] = CStringGetTextDatum(m_plan_table[i]->m_datum->projection->data);
new_record[PT_COST] = Float8GetDatum(m_plan_table[i]->m_datum->cost);
new_record[PT_CARDINALITY] = Float8GetDatum(m_plan_table[i]->m_datum->cardinality);
tuple = (HeapTuple)heap_form_tuple(plan_table_des, new_record, m_plan_table[i]->m_isnull);
* Insert new record into plan_table table
*/
(void)simple_heap_insert(plan_table, tuple);
}
heap_close(plan_table, RowExclusiveLock);
}
* Description: We only control the select/delete permission for pan_table_data.
* For 'plan_table_data' we will handing its permission check espacially, We do not allow ordinary user to
* select/D/U/I
* 1. "Select xx from plan_table" the rte has subquery contain "plan_table_data" and will be pull up in
* subquery_planner.
* 2. In function ExecCheckRTEPerms will check the permission for every rte including those been pulled up.
* 3. So we need to check whether has rte of "plan_table_data" in origin FromClause in this function for select/delete
* cmd. Parameters:
* @ in rangeTable: rte list of query
* Return: return true if there is no rte of 'plan_table_data' in origin FromClause.
*/
bool checkSelectStmtForPlanTable(List* rangeTable)
{
OnlySelectFromPlanTable = false;
ListCell* lc = NULL;
foreach (lc, rangeTable) {
RangeTblEntry* rte = (RangeTblEntry*)lfirst(lc);
if (rte->relname != NULL && strcasecmp(rte->relname, T_PLAN_TABLE_DATA) == 0)
return false;
}
return true;
}
* Description: Deal with 'plan_table_data' permission checking here.
* We do not allow ordinary user to select from 'plan_table_data'.
* Parameters:
* @ in rangeTable: rte of "plan_table_data"
* Return: return -1 for other action(insert/update).
* return 1 if pass permission check.
* return 0 if have no permission.
*/
int checkPermsForPlanTable(RangeTblEntry* rte)
{
* Note that:
* 1. Planner will replace 'plan_table' with 'plan_table_data' in subquery_planner routine.
* If it is only select/delete from plan_table view, we should return ture.
* 2. If current user has superuser or sysdba privileges, we will allow the user do select/delelte on
* 'plan_table_data'.
*/
if (rte->requiredPerms == ACL_SELECT) {
if (OnlySelectFromPlanTable || superuser())
return 1;
else
return 0;
} else if (rte->requiredPerms == (ACL_SELECT + ACL_DELETE)
|| rte->requiredPerms == ACL_DELETE ) {
if (OnlyDeleteFromPlanTable || superuser())
return 1;
else
return 0;
} else {
return -1;
}
}
static inline void appendCSV(StringInfo buf, const char* data)
{
const char* p = data;
char c;
if (p == NULL)
return;
appendStringInfoCharMacro(buf, '"');
while ((c = *p++) != '\0') {
if (c == '"')
appendStringInfoCharMacro(buf, '"');
appendStringInfoCharMacro(buf, c);
}
appendStringInfoCharMacro(buf, '"');
}
* Common code to show sort/group keys, which are represented in plan nodes
* as arrays of targetlist indexes. If it's a sort key rather than a group
* key, also pass sort operators/collations/nullsFirst arrays.
*/
static void show_sort_group_keys(PlanState* planstate, const char* qlabel, int nkeys, const AttrNumber* keycols,
const Oid* sortOperators, const Oid* collations, const bool* nullsFirst, List* ancestors, ExplainState* es)
{
Plan* plan = planstate->plan;
List* context = NIL;
List* result = NIL;
StringInfoData sortkeybuf;
bool useprefix = false;
int keyno;
if (nkeys <= 0)
return;
initStringInfo(&sortkeybuf);
context = deparse_context_for_planstate((Node*)planstate, ancestors, es->rtable);
useprefix = (list_length(es->rtable) > 1 || es->verbose);
for (keyno = 0; keyno < nkeys; keyno++) {
AttrNumber keyresno = keycols[keyno];
TargetEntry* target = get_tle_by_resno(plan->targetlist, keyresno);
* Start/With support
*
* Skip pseudo target entry in SORT list
*/
if (IsPseudoInternalTargetEntry(target)) {
result = lappend(result, pstrdup(target->resname));
continue;
}
char* exprstr = NULL;
if (target == NULL)
ereport(ERROR, (errcode(ERRCODE_INVALID_PARAMETER_VALUE), errmsg("no tlist entry for key %d", keyresno)));
exprstr = deparse_expression((Node*)target->expr, context, useprefix, true);
resetStringInfo(&sortkeybuf);
appendStringInfoString(&sortkeybuf, exprstr);
if (sortOperators != NULL)
show_sortorder_options(
&sortkeybuf, (Node*)target->expr, sortOperators[keyno], collations[keyno], nullsFirst[keyno]);
result = lappend(result, pstrdup(sortkeybuf.data));
}
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "%s: ", qlabel);
ExplainPrettyList(result, es);
} else {
ExplainPropertyList(qlabel, result, es);
}
}
* Append nondefault characteristics of the sort ordering of a column to buf
* (collation, direction, NULLS FIRST/LAST)
*/
static void show_sortorder_options(StringInfo buf, const Node* sortexpr, Oid sortOperator, Oid collation, bool nullsFirst)
{
Oid sortcoltype = exprType(sortexpr);
bool reverse = false;
TypeCacheEntry* typentry = NULL;
typentry = lookup_type_cache(sortcoltype, TYPECACHE_LT_OPR | TYPECACHE_GT_OPR);
* Print COLLATE if it's not default. There are some cases where this is
* redundant, eg if expression is a column whose declared collation is
* that collation, but it's hard to distinguish that here.
*/
if (OidIsValid(collation) && collation != DEFAULT_COLLATION_OID) {
char* collname = get_collation_name(collation);
if (collname == NULL)
ereport(ERROR,
(errcode(ERRCODE_CACHE_LOOKUP_FAILED), errmsg("cache lookup failed for collation %u", collation)));
appendStringInfo(buf, " COLLATE %s", quote_identifier(collname));
}
if (sortOperator == typentry->gt_opr) {
appendStringInfoString(buf, " DESC");
reverse = true;
} else if (sortOperator != typentry->lt_opr) {
char* opname = get_opname(sortOperator);
if (opname == NULL)
ereport(ERROR,
(errcode(ERRCODE_CACHE_LOOKUP_FAILED), errmsg("cache lookup failed for operator %u", sortOperator)));
appendStringInfo(buf, " USING %s", opname);
(void)get_equality_op_for_ordering_op(sortOperator, &reverse);
}
if (nullsFirst && !reverse) {
appendStringInfoString(buf, " NULLS FIRST");
} else if (!nullsFirst && reverse) {
appendStringInfoString(buf, " NULLS LAST");
}
}
* Description: Show tablesample properties.
*
* Parameters:
* @in plan: Scan plan.
* @in planstate: PlanState node.
* @in ancestors: The ancestors list is a list of the PlanState's parent PlanStates.
* @in es: Explain state.
*
* Return: void
*/
static void show_tablesample(Plan* plan, PlanState* planstate, List* ancestors, ExplainState* es)
{
List* context = NIL;
bool useprefix = false;
const char* method_name = NULL;
List* params = NIL;
char* repeatable = NULL;
ListCell* lc = NULL;
TableSampleClause* tsc = ((Scan*)plan)->tablesample;
if (tsc == NULL) {
return;
}
context = deparse_context_for_planstate((Node*)planstate, ancestors, es->rtable);
useprefix = list_length(es->rtable) > 1;
method_name =
(tsc->sampleType == SYSTEM_SAMPLE ? "system" : (tsc->sampleType == BERNOULLI_SAMPLE ? "bernoulli" : "hybrid"));
foreach (lc, tsc->args) {
Node* arg = (Node*)lfirst(lc);
params = lappend(params, deparse_expression(arg, context, useprefix, false));
}
if (tsc->repeatable) {
repeatable = deparse_expression((Node*)tsc->repeatable, context, useprefix, false);
} else {
repeatable = NULL;
}
if (es->format == EXPLAIN_FORMAT_TEXT) {
bool first = true;
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "Sampling: %s (", method_name);
foreach (lc, params) {
if (!first)
appendStringInfoString(es->str, ", ");
appendStringInfoString(es->str, (const char*)lfirst(lc));
first = false;
}
appendStringInfoChar(es->str, ')');
if (repeatable != NULL) {
appendStringInfo(es->str, " REPEATABLE (%s)", repeatable);
}
appendStringInfoChar(es->str, '\n');
} else {
ExplainPropertyText("Sampling Method", method_name, es);
ExplainPropertyList("Sampling Parameters", params, es);
if (repeatable != NULL) {
ExplainPropertyText("Repeatable Seed", repeatable, es);
}
}
list_free_deep(params);
}
void PlanTable::flush_data_to_file()
{
int i = 0;
int j = 0;
int k = 0;
Datum* values = NULL;
bool* nulls = NULL;
bool is_null = false;
Datum data;
Datum val;
Oid typoid;
Oid foutoid;
bool typisvarlena = false;
char* result = NULL;
bool from_datanode = false;
int node_num = 0;
for (i = 0; i < m_plan_size; i++) {
values = m_multi_info[0][i][0].m_Datum;
from_datanode = values[FROM_DATANODE];
node_num = from_datanode ? m_data_size : 1;
for (j = 0; j < node_num; j++) {
values = m_multi_info[j][i][0].m_Datum;
nulls = m_multi_info[j][i][0].m_Nulls;
for (k = 0; k < EXPLAIN_TOTAL_ATTNUM; k++) {
is_null = nulls[k];
data = values[k];
if (is_null) {
if (k == EXPLAIN_TOTAL_ATTNUM - 1)
appendBinaryStringInfo(info_str, "", 0);
else
appendBinaryStringInfo(info_str, ",", 1);
} else {
typoid = m_desc->attrs[k].atttypid;
getTypeOutputInfo(typoid, &foutoid, &typisvarlena);
if (typisvarlena)
val = PointerGetDatum(PG_DETOAST_DATUM(data));
else
val = data;
result = OidOutputFunctionCall(foutoid, val);
if (val != data)
pfree(DatumGetPointer(val));
if (typisvarlena)
appendCSV(info_str, result);
else if (typoid == BOOLOID) {
if (strcmp(result, "t") == 0)
appendStringInfo(info_str, "true");
else
appendStringInfo(info_str, "false");
} else
appendStringInfo(info_str, "%s", result);
if (k != EXPLAIN_TOTAL_ATTNUM - 1)
appendStringInfoChar(info_str, ',');
}
}
appendStringInfoChar(info_str, '\n');
}
}
}
* Insert obs scan information into pg_obsscaninfo table
*/
void insert_obsscaninfo(
uint64 queryid, const char* rel_name, int64 file_count, double scan_data_size, double total_time, int format)
{
Relation pgobsscaninforel;
HeapTuple htup;
bool nulls[Natts_pg_obsscaninfo];
Datum values[Natts_pg_obsscaninfo];
int i;
TimestampTz ts = GetCurrentTimestamp();
StringInfo insert_stmt1, insert_stmt2;
const char* billing_info = "unbilled";
char* file_format = NULL;
if (format == 0)
file_format = "orc";
else if (format == 1)
file_format = "text";
else if (format == 2)
file_format = "csv";
else
file_format = "unknown";
MemoryContext current_ctx = CurrentMemoryContext;
for (i = 0; i < Natts_pg_obsscaninfo; i++) {
nulls[i] = false;
values[i] = (Datum)0;
}
values[Anum_pg_obsscaninfo_query_id - 1] = Int64GetDatum(queryid);
values[Anum_pg_obsscaninfo_user_id - 1] = CStringGetTextDatum(u_sess->misc_cxt.CurrentUserName);
values[Anum_pg_obsscaninfo_table_name - 1] = CStringGetTextDatum(rel_name);
values[Anum_pg_obsscaninfo_file_type - 1] = CStringGetTextDatum(file_format);
values[Anum_pg_obsscaninfo_time_stamp - 1] = TimestampTzGetDatum(ts);
values[Anum_pg_obsscaninfo_actual_time - 1] = Float8GetDatum(total_time);
values[Anum_pg_obsscaninfo_file_scanned - 1] = file_count;
values[Anum_pg_obsscaninfo_data_size - 1] = Float8GetDatum(scan_data_size);
values[Anum_pg_obsscaninfo_billing_info - 1] = CStringGetTextDatum(billing_info);
pgobsscaninforel = heap_open(ObsScanInfoRelationId, RowExclusiveLock);
htup = heap_form_tuple(pgobsscaninforel->rd_att, values, nulls);
(void)simple_heap_insert(pgobsscaninforel, htup);
CatalogUpdateIndexes(pgobsscaninforel, htup);
tableam_tops_free_tuple(htup);
heap_close(pgobsscaninforel, RowExclusiveLock);
pgobsscaninforel = heap_open(GSObsScanInfoRelationId, RowExclusiveLock);
htup = heap_form_tuple(pgobsscaninforel->rd_att, values, nulls);
(void)simple_heap_insert(pgobsscaninforel, htup);
CatalogUpdateIndexes(pgobsscaninforel, htup);
tableam_tops_free_tuple(htup);
heap_close(pgobsscaninforel, RowExclusiveLock);
elog(DEBUG1, "SyncUp with other CN");
insert_stmt1 = makeStringInfo();
appendStringInfo(insert_stmt1,
"insert into pg_obsscaninfo values (%lu, '%s', '%s', '%s', '%s', '%lf', %ld, %lf, '%s');",
queryid,
u_sess->misc_cxt.CurrentUserName,
rel_name,
file_format,
timestamptz_to_str(ts),
total_time,
file_count,
scan_data_size,
billing_info);
insert_stmt2 = makeStringInfo();
appendStringInfo(insert_stmt2,
"insert into gs_obsscaninfo values (%lu, '%s', '%s', '%s', '%s', '%lf', %ld, %lf, '%s');",
queryid,
u_sess->misc_cxt.CurrentUserName,
rel_name,
file_format,
timestamptz_to_str(ts),
total_time,
file_count,
scan_data_size,
billing_info);
elog(DEBUG1, "%s", insert_stmt1->data);
elog(DEBUG1, "%s", insert_stmt2->data);
PG_TRY();
{
* If insert job status in local success and only synchronize to other coordinator fail,
* we should consider the worker success, so that we at least have one record for billing.
*/
ExecUtilityStmtOnNodes(insert_stmt1->data, NULL, false, false, EXEC_ON_COORDS, false);
ExecUtilityStmtOnNodes(insert_stmt2->data, NULL, false, false, EXEC_ON_COORDS, false);
}
PG_CATCH();
{
* So if we invoke LWLockrelease, we must call HOLD_INTERRUPT
* the function called by OBSInstrumentation::insertData() which holds
* OBSRuntimeLock.
*/
HOLD_INTERRUPTS();
MemoryContextSwitchTo(current_ctx);
ErrorData* edata = CopyErrorData();
FlushErrorState();
ereport(LOG,
(errcode(ERRCODE_OPERATE_FAILED),
errmsg("Synchronize OBS scan info to other coordinator failed, OBS billing info may be out of sync"),
errdetail("Synchronize fail reason: %s.", edata->message)));
FreeErrorData(edata);
}
PG_END_TRY();
pfree_ext(insert_stmt1);
pfree_ext(insert_stmt2);
}
static void show_unique_check_info(PlanState *planstate, ExplainState *es)
{
Agg *aggnode = (Agg *) planstate->plan;
if (aggnode->unique_check) {
const char *str = "Unique Check Required";
if (t_thrd.explain_cxt.explain_perf_mode != EXPLAIN_NORMAL && es->planinfo->m_verboseInfo) {
es->planinfo->m_verboseInfo->set_plan_name<false, true>();
appendStringInfo(es->planinfo->m_verboseInfo->info_str, "%s\n", str);
}
else {
if (es->format == EXPLAIN_FORMAT_TEXT && !es->wlm_statistics_plan_max_digit) {
appendStringInfoSpaces(es->str, es->indent * 2);
appendStringInfo(es->str, "%s\n", str);
}
}
}
}
static void show_ndpplugin_statistic(ExplainState *es, PlanState* planstate, StringInfo str, bool is_pretty)
{
Plan* plan = planstate->plan;
if (!plan->ndp_pushdown_optimized &&
!(plan->lefttree && plan->type == T_Agg && plan->lefttree->ndp_pushdown_optimized &&
reinterpret_cast<NdpScanCondition*>(plan->lefttree->ndp_pushdown_condition)->plan == plan)) {
return;
}
TableScanDesc desc = reinterpret_cast<ScanState*>(planstate)->ss_currentScanDesc;
if (desc && !desc->ndp_pushdown_optimized) {
return;
}
if (is_pretty) {
appendStringInfo(str, " NDPpushdown");
} else {
appendStringInfo(es->str, " NDPpushdown");
}
if (!es->analyze) {
return;
}
Instrumentation* instr;
int pushdownPage = 0;
int normalPage = 0;
int ndpPage = 0;
int dop = planstate->plan->parallel_enabled ? planstate->plan->dop : 1;
if (planstate->plan->plan_node_id > 0 && u_sess->instr_cxt.global_instr &&
u_sess->instr_cxt.global_instr->isFromDataNode(planstate->plan->plan_node_id)) {
for (int i = 0; i < u_sess->instr_cxt.global_instr->getInstruNodeNum(); i++) {
ThreadInstrumentation* threadinstr =
u_sess->instr_cxt.global_instr->getThreadInstrumentation(i, planstate->plan->plan_node_id, 0);
if (threadinstr == NULL)
continue;
for (int j = 0; j < dop; j++) {
instr = u_sess->instr_cxt.global_instr->getInstrSlot(i, planstate->plan->plan_node_id, j);
if (instr != NULL && instr->nloops > 0) {
pushdownPage += instr->ndp_pushdown_page;
normalPage += instr->ndp_sendback_page;
ndpPage += instr->ndp_handled;
}
}
}
}
instr = planstate->instrument;
pushdownPage += instr->ndp_pushdown_page;
normalPage += instr->ndp_sendback_page;
ndpPage += instr->ndp_handled;
appendStringInfo(es->str, " (total page: %d, back to normal page: %d, ndp handled: %d)", pushdownPage, normalPage, ndpPage);
}
void ExplainDatumProperty(char const *name, Datum const value, Oid const type, ExplainState* es)
{
Datum output_datum = 0;
char const *output = nullptr;
if (type_is_array_domain(type)) {
output_datum = OidFunctionCall2(F_ARRAY_OUT, value, type);
output = DatumGetCString(output_datum);
ExplainPropertyText(name, output, es);
} else {
switch (type) {
case INT1OID:
ExplainPropertyInteger(name, DatumGetInt8(value), es);
break;
case INT2OID:
ExplainPropertyInteger(name, DatumGetInt16(value), es);
break;
case INT4OID:
ExplainPropertyInteger(name, DatumGetInt32(value), es);
break;
case INT8OID:
ExplainPropertyLong(name, DatumGetInt64(value), es);
break;
case FLOAT4OID:
ExplainPropertyFloat(name, DatumGetFloat4(value), 6, es);
break;
case FLOAT8OID:
ExplainPropertyFloat(name, DatumGetFloat8(value), 10, es);
break;
default:
output = Datum_to_string(value, type, type == InvalidOid);
ExplainPropertyText(name, output, es);
break;
}
}
}
* this function explains a group of model information (recursively if necessary)
* as provided by the list
*/
ListCell* ExplainTrainInfoGroup(List const* group, ExplainState *es, bool const group_opened)
{
ListCell *lc = nullptr;
TrainingInfo *train_info = nullptr;
* in a single traversal of the group there will (potentially) be
* recursive calls that will print nested sub-groups (by properly
* opening and closing them)
*/
foreach(lc, group) {
train_info = lfirst_node(TrainingInfo, lc);
if (train_info->open_group) {
* if the cell for some reason has both opening and closing booleans set we ignore it
*/
if (train_info->close_group)
continue;
if (!train_info->name)
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_SYNTAX_ERROR),
errmsg("the name of a group cannot be empty (when opening)")));
ExplainOpenGroup(train_info->name, train_info->name, true, es);
List next_group;
next_group.length = 0;
next_group.type = group->type;
next_group.head = lnext(lc);
next_group.tail = group->tail;
* once we have printed a whole group we gotta fast-forward to the first element after the group.
* the returned ListCell is the closing element of the group we just printed (foreach will do the rest)
*/
lc = ExplainTrainInfoGroup(&next_group, es, true);
* if there are more opening groups than closing one at some point the list will have to get
* exhausted and there will be recursive call that cannot cleanly finish
*/
if (!lc)
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_SYNTAX_ERROR),
errmsg("group of name \"%s\" was not closed", train_info->name)));
auto current_train_info = lfirst_node(TrainingInfo, lc);
* the name of opening and closing groups has to match, otherwise error.
* observe that both strings are non-empty (otherwise error before this)
*/
if (strcmp(train_info->name, current_train_info->name) != 0)
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_SYNTAX_ERROR),
errmsg("group of name \"%s\" was not properly closed, closing group name was \"%s\"",
train_info->name, current_train_info->name)));
continue;
} else if (train_info->close_group) {
* this is in charge of validating that closing a group is correct (by having opened at the
* parent call)
*/
if (!group_opened)
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_SYNTAX_ERROR),
errmsg("badly formed group of name \"%s\" of training information (opening group not found)",
train_info->name)));
if (!train_info->name)
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_SYNTAX_ERROR),
errmsg("the name of a group cannot be empty (when closing)")));
ExplainCloseGroup(train_info->name, train_info->name, true, es);
break;
} else {
* we output the next object of the group
*/
ExplainDatumProperty(train_info->name, train_info->value, train_info->type, es);
}
}
return lc;
}
void do_model_explain(ExplainState *es, const Model *model)
{
ListCell *lc = nullptr;
ExplainBeginOutput(es);
* there is a common part of each model that can be formatted now
*/
ExplainOpenGroup("General information", "General information", true, es);
ExplainPropertyText("Name", model->model_name, es);
ExplainPropertyText("Algorithm", algorithm_ml_to_string(model->algorithm), es);
if (model->sql != NULL)
ExplainPropertyText("Query", model->sql, es);
char const *return_type_str = nullptr;
switch (model->return_type) {
case BOOLOID:
return_type_str = prediction_type_to_string(TYPE_BOOL);
break;
case BYTEAOID:
return_type_str = prediction_type_to_string(TYPE_BYTEA);
break;
case INT4OID:
return_type_str = prediction_type_to_string(TYPE_INT32);
break;
case INT8OID:
return_type_str = prediction_type_to_string(TYPE_INT64);
break;
case FLOAT4OID:
return_type_str = prediction_type_to_string(TYPE_FLOAT32);
break;
case FLOAT8OID:
return_type_str = prediction_type_to_string(TYPE_FLOAT64);
break;
case FLOAT8ARRAYOID:
return_type_str = prediction_type_to_string(TYPE_FLOAT64ARRAY);
break;
case NUMERICOID:
return_type_str = prediction_type_to_string(TYPE_NUMERIC);
break;
case TEXTOID:
return_type_str = prediction_type_to_string(TYPE_TEXT);
break;
case VARCHAROID:
return_type_str = prediction_type_to_string(TYPE_VARCHAR);
break;
default:
ereport(ERROR, (errmodule(MOD_DB4AI), errcode(ERRCODE_INVALID_PARAMETER_VALUE),
errmsg("invalid return type (%d) for model of name \"%s\"", model->return_type,
model->model_name)));
}
ExplainPropertyText("Return type", return_type_str, es);
ExplainPropertyFloat("Pre-processing time", model->pre_time_secs, 6, es);
ExplainPropertyFloat("Execution time", model->exec_time_secs, 6, es);
ExplainPropertyInteger("Processed tuples", model->processed_tuples, es);
ExplainPropertyInteger("Discarded tuples", model->discarded_tuples, es);
ExplainCloseGroup("General information", "General information", true, es);
* hyper-parameters are now output
*/
ExplainOpenGroup("Hyper-parameters", "Hyper-parameters", true, es);
foreach(lc, model->hyperparameters) {
auto hyperparameter = lfirst_node(Hyperparameter, lc);
ExplainDatumProperty(hyperparameter->name, hyperparameter->value, hyperparameter->type, es);
}
ExplainCloseGroup("Hyper-parameters", "Hyper-parameters", true, es);
* scores are now output
*/
ExplainOpenGroup("Scores", "Scores", true, es);
foreach(lc, model->scores) {
auto training_score = lfirst_node(TrainingScore, lc);
ExplainPropertyFloat(training_score->name, training_score->value, 10, es);
}
ExplainCloseGroup("Scores", "Scores", true, es);
* Model-dependent information
*/
AlgorithmAPI *algo_api = get_algorithm_api(model->algorithm);
if (algo_api->explain && model->data.version > DB4AI_MODEL_V00) {
List *train_data = algo_api->explain(algo_api, &model->data, model->return_type);
if (train_data) {
ExplainOpenGroup("Train data", "Train data", true, es);
ExplainTrainInfoGroup(train_data, es, false);
ExplainCloseGroup("Train data", "Train data", true, es);
}
}
ExplainEndOutput(es);
}
