0001-show-stats-in-explain-rebased.patch

application/octet-stream

Filename: 0001-show-stats-in-explain-rebased.patch
Type: application/octet-stream
Part: 0
Message: Re: Showing applied extended statistics in explain

Patch

Format: unified
Series: patch 0001
File+
src/backend/commands/explain.c 97 0
src/backend/nodes/copyfuncs.c 4 0
src/backend/nodes/makefuncs.c 11 0
src/backend/nodes/outfuncs.c 4 0
src/backend/nodes/readfuncs.c 4 0
src/backend/optimizer/plan/createplan.c 15 0
src/backend/optimizer/util/relnode.c 12 0
src/backend/optimizer/util/restrictinfo.c 35 0
src/backend/statistics/extended_stats.c 5 0
src/backend/utils/adt/selfuncs.c 11 0
src/backend/utils/cache/lsyscache.c 49 0
src/include/nodes/makefuncs.h 2 0
src/include/nodes/pathnodes.h 5 0
src/include/nodes/plannodes.h 5 0
src/include/optimizer/restrictinfo.h 2 0
src/include/utils/lsyscache.h 3 0
diff --git a/src/backend/commands/explain.c b/src/backend/commands/explain.c
index b970997..b351220 100644
--- a/src/backend/commands/explain.c
+++ b/src/backend/commands/explain.c
@@ -78,6 +78,9 @@ static void show_qual(List *qual, const char *qlabel,
 static void show_scan_qual(List *qual, const char *qlabel,
 						   PlanState *planstate, List *ancestors,
 						   ExplainState *es);
+static void show_scan_stats(List *stats, List *clauses, List *ors,
+							PlanState *planstate, List *ancestors,
+							ExplainState *es);
 static void show_upper_qual(List *qual, const char *qlabel,
 							PlanState *planstate, List *ancestors,
 							ExplainState *es);
@@ -1742,6 +1745,10 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			if (plan->qual)
 				show_instrumentation_count("Rows Removed by Filter", 1,
 										   planstate, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_IndexOnlyScan:
 			show_scan_qual(((IndexOnlyScan *) plan)->indexqual,
@@ -1758,10 +1765,18 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			if (es->analyze)
 				ExplainPropertyFloat("Heap Fetches", NULL,
 									 planstate->instrument->ntuples2, 0, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_BitmapIndexScan:
 			show_scan_qual(((BitmapIndexScan *) plan)->indexqualorig,
 						   "Index Cond", planstate, ancestors, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_BitmapHeapScan:
 			show_scan_qual(((BitmapHeapScan *) plan)->bitmapqualorig,
@@ -1791,6 +1806,10 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			if (plan->qual)
 				show_instrumentation_count("Rows Removed by Filter", 1,
 										   planstate, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_Gather:
 			{
@@ -1978,6 +1997,10 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			if (plan->qual)
 				show_instrumentation_count("Rows Removed by Filter", 1,
 										   planstate, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_Group:
 			show_group_keys(castNode(GroupState, planstate), ancestors, es);
@@ -1985,6 +2008,10 @@ ExplainNode(PlanState *planstate, List *ancestors,
 			if (plan->qual)
 				show_instrumentation_count("Rows Removed by Filter", 1,
 										   planstate, es);
+			if (es->verbose)
+				show_scan_stats(plan->applied_stats, plan->applied_clauses,
+								plan->applied_clauses_or,
+								planstate, ancestors, es);
 			break;
 		case T_Sort:
 			show_sort_keys(castNode(SortState, planstate), ancestors, es);
@@ -2305,6 +2332,76 @@ show_scan_qual(List *qual, const char *qlabel,
 }
 
 /*
+ * Show a generic expression
+ */
+static char *
+deparse_stat_expression(Node *node,
+						PlanState *planstate, List *ancestors,
+						ExplainState *es)
+{
+	List	   *context;
+
+	/* Set up deparsing context */
+	context = set_deparse_context_plan(es->deparse_cxt,
+									   planstate->plan,
+									   ancestors);
+
+	/* Deparse the expression */
+	return deparse_expression(node, context, false, false);
+}
+
+/*
+ * Show a qualifier expression (which is a List with implicit AND semantics)
+ */
+static char *
+show_stat_qual(List *qual, bool is_or,
+			   PlanState *planstate, List *ancestors,
+			   ExplainState *es)
+{
+	Node	   *node;
+
+	/* No work if empty qual */
+	if (qual == NIL)
+		return NULL;
+
+	/* Convert AND list to explicit AND */
+	if (is_or)
+		node = (Node *) make_ors_explicit(qual);
+	else
+		node = (Node *) make_ands_explicit(qual);
+
+	/* And show it */
+	return deparse_stat_expression(node, planstate, ancestors, es);
+}
+
+/*
+ * Show applied statistics for scan plan node
+ */
+static void
+show_scan_stats(List *stats, List *clauses, List *ors,
+				PlanState *planstate, List *ancestors, ExplainState *es)
+{
+	ListCell   *lc1, *lc2, *lc3;
+	StringInfoData	str;
+
+	forthree (lc1, stats, lc2, clauses, lc3, ors)
+	{
+		StatisticExtInfo   *stat = (StatisticExtInfo *) lfirst(lc1);
+		List   *clauses = (List *) lfirst(lc2);
+		int		is_or = lfirst_int(lc3);
+
+		initStringInfo(&str);
+
+		appendStringInfo(&str, "%s.%s  Clauses: %s",
+						 get_namespace_name(get_statistics_namespace(stat->statOid)),
+						 get_statistics_name(stat->statOid),
+						 show_stat_qual(clauses, is_or, planstate, ancestors, es));
+
+		ExplainPropertyText("Statistics", str.data, es);
+	}
+}
+
+/*
  * Show a qualifier expression for an upper-level plan node
  */
 static void
diff --git a/src/backend/nodes/copyfuncs.c b/src/backend/nodes/copyfuncs.c
index 90b5da5..654c4b5 100644
--- a/src/backend/nodes/copyfuncs.c
+++ b/src/backend/nodes/copyfuncs.c
@@ -136,6 +136,10 @@ CopyPlanFields(const Plan *from, Plan *newnode)
 	COPY_NODE_FIELD(initPlan);
 	COPY_BITMAPSET_FIELD(extParam);
 	COPY_BITMAPSET_FIELD(allParam);
+
+	COPY_NODE_FIELD(applied_stats);
+	COPY_NODE_FIELD(applied_clauses);
+	COPY_NODE_FIELD(applied_clauses_or);
 }
 
 /*
diff --git a/src/backend/nodes/makefuncs.c b/src/backend/nodes/makefuncs.c
index 8223956..26b9255 100644
--- a/src/backend/nodes/makefuncs.c
+++ b/src/backend/nodes/makefuncs.c
@@ -715,6 +715,17 @@ make_ands_explicit(List *andclauses)
 		return make_andclause(andclauses);
 }
 
+Expr *
+make_ors_explicit(List *orclauses)
+{
+	if (orclauses == NIL)
+		return (Expr *) makeBoolConst(true, false);
+	else if (list_length(orclauses) == 1)
+		return (Expr *) linitial(orclauses);
+	else
+		return make_orclause(orclauses);
+}
+
 List *
 make_ands_implicit(Expr *clause)
 {
diff --git a/src/backend/nodes/outfuncs.c b/src/backend/nodes/outfuncs.c
index 2b02369..6a6e20c 100644
--- a/src/backend/nodes/outfuncs.c
+++ b/src/backend/nodes/outfuncs.c
@@ -349,6 +349,10 @@ _outPlanInfo(StringInfo str, const Plan *node)
 	WRITE_NODE_FIELD(initPlan);
 	WRITE_BITMAPSET_FIELD(extParam);
 	WRITE_BITMAPSET_FIELD(allParam);
+
+	WRITE_NODE_FIELD(applied_stats);
+	WRITE_NODE_FIELD(applied_clauses);
+	WRITE_NODE_FIELD(applied_clauses_or);
 }
 
 /*
diff --git a/src/backend/nodes/readfuncs.c b/src/backend/nodes/readfuncs.c
index 3f68f7c..7220948 100644
--- a/src/backend/nodes/readfuncs.c
+++ b/src/backend/nodes/readfuncs.c
@@ -1627,6 +1627,10 @@ ReadCommonPlan(Plan *local_node)
 	READ_NODE_FIELD(initPlan);
 	READ_BITMAPSET_FIELD(extParam);
 	READ_BITMAPSET_FIELD(allParam);
+
+	READ_NODE_FIELD(applied_stats);
+	READ_NODE_FIELD(applied_clauses);
+	READ_NODE_FIELD(applied_clauses_or);
 }
 
 /*
diff --git a/src/backend/optimizer/plan/createplan.c b/src/backend/optimizer/plan/createplan.c
index cd6d72c..9efb961 100644
--- a/src/backend/optimizer/plan/createplan.c
+++ b/src/backend/optimizer/plan/createplan.c
@@ -5294,12 +5294,27 @@ order_qual_clauses(PlannerInfo *root, List *clauses)
 static void
 copy_generic_path_info(Plan *dest, Path *src)
 {
+	ListCell *lc;
+
 	dest->startup_cost = src->startup_cost;
 	dest->total_cost = src->total_cost;
 	dest->plan_rows = src->rows;
 	dest->plan_width = src->pathtarget->width;
 	dest->parallel_aware = src->parallel_aware;
 	dest->parallel_safe = src->parallel_safe;
+
+	dest->applied_stats = src->parent->applied_stats;
+	dest->applied_clauses_or = src->parent->applied_clauses_or;
+
+	dest->applied_clauses = NIL;
+	foreach (lc, src->parent->applied_clauses)
+	{
+		List *clauses = (List *) lfirst(lc);
+
+		dest->applied_clauses
+			= lappend(dest->applied_clauses,
+					  maybe_extract_actual_clauses(clauses, false));
+	}
 }
 
 /*
diff --git a/src/backend/optimizer/util/relnode.c b/src/backend/optimizer/util/relnode.c
index 520409f..6a4e1c1 100644
--- a/src/backend/optimizer/util/relnode.c
+++ b/src/backend/optimizer/util/relnode.c
@@ -260,6 +260,10 @@ build_simple_rel(PlannerInfo *root, int relid, RelOptInfo *parent)
 	rel->partexprs = NULL;
 	rel->nullable_partexprs = NULL;
 
+	rel->applied_stats = NIL;
+	rel->applied_clauses = NIL;
+	rel->applied_clauses_or = NIL;
+
 	/*
 	 * Pass assorted information down the inheritance hierarchy.
 	 */
@@ -675,6 +679,10 @@ build_join_rel(PlannerInfo *root,
 	joinrel->partexprs = NULL;
 	joinrel->nullable_partexprs = NULL;
 
+	joinrel->applied_stats = NIL;
+	joinrel->applied_clauses = NIL;
+	joinrel->applied_clauses_or = NIL;
+
 	/* Compute information relevant to the foreign relations. */
 	set_foreign_rel_properties(joinrel, outer_rel, inner_rel);
 
@@ -854,6 +862,10 @@ build_child_join_rel(PlannerInfo *root, RelOptInfo *outer_rel,
 	joinrel->partexprs = NULL;
 	joinrel->nullable_partexprs = NULL;
 
+	joinrel->applied_stats = NIL;
+	joinrel->applied_clauses = NIL;
+	joinrel->applied_clauses_or = NIL;
+
 	joinrel->top_parent_relids = bms_union(outer_rel->top_parent_relids,
 										   inner_rel->top_parent_relids);
 
diff --git a/src/backend/optimizer/util/restrictinfo.c b/src/backend/optimizer/util/restrictinfo.c
index ef8df3d..cd982c3 100644
--- a/src/backend/optimizer/util/restrictinfo.c
+++ b/src/backend/optimizer/util/restrictinfo.c
@@ -459,6 +459,41 @@ extract_actual_clauses(List *restrictinfo_list,
 }
 
 /*
+ * maybe_extract_actual_clauses
+ *
+ * Just like extract_actual_clauses, but does not require the clauses to
+ * already be RestrictInfo.
+ *
+ * XXX Does not handle RestrictInfos nested in OR clauses.
+ */
+List *
+maybe_extract_actual_clauses(List *restrictinfo_list,
+							 bool pseudoconstant)
+{
+	List	   *result = NIL;
+	ListCell   *l;
+
+	foreach(l, restrictinfo_list)
+	{
+		RestrictInfo   *rinfo;
+		Node *node = (Node *) lfirst(l);
+
+		if (!IsA(node, RestrictInfo))
+		{
+			result = lappend(result, node);
+			continue;
+		}
+
+		rinfo = (RestrictInfo *) node;
+
+		if (rinfo->pseudoconstant == pseudoconstant)
+			result = lappend(result, rinfo->clause);
+	}
+
+	return result;
+}
+
+/*
  * extract_actual_join_clauses
  *
  * Extract bare clauses from 'restrictinfo_list', separating those that
diff --git a/src/backend/statistics/extended_stats.c b/src/backend/statistics/extended_stats.c
index 87fe82e..3606e35 100644
--- a/src/backend/statistics/extended_stats.c
+++ b/src/backend/statistics/extended_stats.c
@@ -1829,6 +1829,11 @@ statext_mcv_clauselist_selectivity(PlannerInfo *root, List *clauses, int varReli
 			list_exprs[listidx] = NULL;
 		}
 
+		/* add it to the list of applied stats/clauses */
+		rel->applied_stats = lappend(rel->applied_stats, stat);
+		rel->applied_clauses = lappend(rel->applied_clauses, stat_clauses);
+		rel->applied_clauses_or = lappend_int(rel->applied_clauses_or, (is_or) ? 1 : 0);
+
 		if (is_or)
 		{
 			bool	   *or_matches = NULL;
diff --git a/src/backend/utils/adt/selfuncs.c b/src/backend/utils/adt/selfuncs.c
index 1fbb0b2..c03a2e3 100644
--- a/src/backend/utils/adt/selfuncs.c
+++ b/src/backend/utils/adt/selfuncs.c
@@ -4019,6 +4019,7 @@ estimate_multivariate_ndistinct(PlannerInfo *root, RelOptInfo *rel,
 		ListCell   *lc2;
 		Bitmapset  *matched = NULL;
 		AttrNumber	attnum_offset;
+		List	   *matched_exprs = NIL;
 
 		/*
 		 * How much we need to offset the attnums? If there are no
@@ -4066,6 +4067,9 @@ estimate_multivariate_ndistinct(PlannerInfo *root, RelOptInfo *rel,
 
 				matched = bms_add_member(matched, attnum);
 
+				/* track expressions matched by this statistics */
+				matched_exprs = lappend(matched_exprs, varinfo->var);
+
 				found = true;
 			}
 
@@ -4094,6 +4098,9 @@ estimate_multivariate_ndistinct(PlannerInfo *root, RelOptInfo *rel,
 
 					matched = bms_add_member(matched, attnum);
 
+					/* track expressions matched by this statistics */
+					matched_exprs = lappend(matched_exprs, expr);
+
 					/* there should be just one matching expression */
 					break;
 				}
@@ -4102,6 +4109,10 @@ estimate_multivariate_ndistinct(PlannerInfo *root, RelOptInfo *rel,
 			}
 		}
 
+		rel->applied_stats = lappend(rel->applied_stats, matched_info);
+		rel->applied_clauses = lappend(rel->applied_clauses, matched_exprs);
+		rel->applied_clauses_or = lappend(rel->applied_clauses_or, 0);
+
 		/* Find the specific item that exactly matches the combination */
 		for (i = 0; i < stats->nitems; i++)
 		{
diff --git a/src/backend/utils/cache/lsyscache.c b/src/backend/utils/cache/lsyscache.c
index feef999..3363be4 100644
--- a/src/backend/utils/cache/lsyscache.c
+++ b/src/backend/utils/cache/lsyscache.c
@@ -33,6 +33,7 @@
 #include "catalog/pg_proc.h"
 #include "catalog/pg_range.h"
 #include "catalog/pg_statistic.h"
+#include "catalog/pg_statistic_ext.h"
 #include "catalog/pg_transform.h"
 #include "catalog/pg_type.h"
 #include "miscadmin.h"
@@ -3568,3 +3569,51 @@ get_index_isclustered(Oid index_oid)
 
 	return isclustered;
 }
+
+/*
+ * get_statistics_name
+ *		Returns the name of a given extended statistics
+ *
+ * Returns a palloc'd copy of the string, or NULL if no such namespace.
+ */
+char *
+get_statistics_name(Oid stxid)
+{
+	HeapTuple	tp;
+
+	tp = SearchSysCache1(STATEXTOID, ObjectIdGetDatum(stxid));
+	if (HeapTupleIsValid(tp))
+	{
+		Form_pg_statistic_ext stxtup = (Form_pg_statistic_ext) GETSTRUCT(tp);
+		char	   *result;
+
+		result = pstrdup(NameStr(stxtup->stxname));
+		ReleaseSysCache(tp);
+		return result;
+	}
+	else
+		return NULL;
+}
+
+/*
+ * get_statistics_namespace
+ *		Returns the namespace OID of a given extended statistics
+ */
+Oid
+get_statistics_namespace(Oid stxid)
+{
+	HeapTuple	tp;
+
+	tp = SearchSysCache1(STATEXTOID, ObjectIdGetDatum(stxid));
+	if (HeapTupleIsValid(tp))
+	{
+		Form_pg_statistic_ext stxtup = (Form_pg_statistic_ext) GETSTRUCT(tp);
+		Oid		result;
+
+		result = stxtup->stxnamespace;
+		ReleaseSysCache(tp);
+		return result;
+	}
+	else
+		return InvalidOid;
+}
diff --git a/src/include/nodes/makefuncs.h b/src/include/nodes/makefuncs.h
index fe17310..249833d 100644
--- a/src/include/nodes/makefuncs.h
+++ b/src/include/nodes/makefuncs.h
@@ -94,6 +94,8 @@ extern Node *make_and_qual(Node *qual1, Node *qual2);
 extern Expr *make_ands_explicit(List *andclauses);
 extern List *make_ands_implicit(Expr *clause);
 
+extern Expr *make_ors_explicit(List *orclauses);
+
 extern IndexInfo *makeIndexInfo(int numattrs, int numkeyattrs, Oid amoid,
 								List *expressions, List *predicates,
 								bool unique, bool isready, bool concurrent);
diff --git a/src/include/nodes/pathnodes.h b/src/include/nodes/pathnodes.h
index 1f3845b..6abfea6 100644
--- a/src/include/nodes/pathnodes.h
+++ b/src/include/nodes/pathnodes.h
@@ -773,6 +773,11 @@ typedef struct RelOptInfo
 	Relids		all_partrels;	/* Relids set of all partition relids */
 	List	  **partexprs;		/* Non-nullable partition key expressions */
 	List	  **nullable_partexprs; /* Nullable partition key expressions */
+
+	/* info about applied extended statistics */
+	List	   *applied_stats;		/* list of StatisticExtInfo */
+	List	   *applied_clauses;	/* list of lists (of clauses) */
+	List	   *applied_clauses_or;	/* are the clauses AND or OR */
 } RelOptInfo;
 
 /*
diff --git a/src/include/nodes/plannodes.h b/src/include/nodes/plannodes.h
index 0b518ce..425d659 100644
--- a/src/include/nodes/plannodes.h
+++ b/src/include/nodes/plannodes.h
@@ -158,6 +158,11 @@ typedef struct Plan
 	 */
 	Bitmapset  *extParam;
 	Bitmapset  *allParam;
+
+	/* info about applied statistics */
+	List	   *applied_stats;
+	List	   *applied_clauses;
+	List	   *applied_clauses_or;
 } Plan;
 
 /* ----------------
diff --git a/src/include/optimizer/restrictinfo.h b/src/include/optimizer/restrictinfo.h
index 6d30bd5..fc27723 100644
--- a/src/include/optimizer/restrictinfo.h
+++ b/src/include/optimizer/restrictinfo.h
@@ -37,6 +37,8 @@ extern bool restriction_is_securely_promotable(RestrictInfo *restrictinfo,
 extern List *get_actual_clauses(List *restrictinfo_list);
 extern List *extract_actual_clauses(List *restrictinfo_list,
 									bool pseudoconstant);
+extern List *maybe_extract_actual_clauses(List *restrictinfo_list,
+										  bool pseudoconstant);
 extern void extract_actual_join_clauses(List *restrictinfo_list,
 										Relids joinrelids,
 										List **joinquals,
diff --git a/src/include/utils/lsyscache.h b/src/include/utils/lsyscache.h
index b8dd27d..a9c206a 100644
--- a/src/include/utils/lsyscache.h
+++ b/src/include/utils/lsyscache.h
@@ -199,6 +199,9 @@ extern bool get_index_isreplident(Oid index_oid);
 extern bool get_index_isvalid(Oid index_oid);
 extern bool get_index_isclustered(Oid index_oid);
 
+extern char *get_statistics_name(Oid stxid);
+extern Oid	get_statistics_namespace(Oid stxid);
+
 #define type_is_array(typid)  (get_element_type(typid) != InvalidOid)
 /* type_is_array_domain accepts both plain arrays and domains over arrays */
 #define type_is_array_domain(typid)  (get_base_element_type(typid) != InvalidOid)