Use ExprStates for hashing in GROUP BY and SubPlans

This speeds up obtaining hash values for GROUP BY and hashed SubPlans by
using the ExprState support for hashing, thus allowing JIT compilation for
obtaining hash values for these operations.

This, even without JIT compilation, has been shown to improve Hash
Aggregate performance in some cases by around 15% and hashed NOT IN
queries in one case by over 30%, however, real-world cases are likely to
see smaller gains as the test cases used were purposefully designed to
have high hashing overheads by keeping the hash table small to prevent
additional memory overheads that would be a factor when working with large
hash tables.

In passing, fix a hypothetical bug in ExecBuildHash32Expr() so that the
initial value is stored directly in the ExprState's result field if
there are no expressions to hash.  None of the current users of this
function use an initial value, so the bug is only hypothetical.

Reviewed-by: Andrei Lepikhov <lepihov@gmail.com>
Discussion: https://postgr.es/m/CAApHDvpYSO3kc9UryMevWqthTBrxgfd9djiAjKHMPUSQeX9vdQ@mail.gmail.com

1 files changed, 14 insertions, 4 deletions

diff --git a/src/backend/executor/nodeSubplan.c b/src/backend/executor/nodeSubplan.c
index 236222d72a1..f26c883c67d 100644
--- a/src/backend/executor/nodeSubplan.c
+++ b/src/backend/executor/nodeSubplan.c
@@ -160,7 +160,7 @@ ExecHashSubPlan(SubPlanState *node,
 			FindTupleHashEntry(node->hashtable,
 							   slot,
 							   node->cur_eq_comp,
-							   node->lhs_hash_funcs) != NULL)
+							   node->lhs_hash_expr) != NULL)
 		{
 			ExecClearTuple(slot);
 			return BoolGetDatum(true);
@@ -857,7 +857,6 @@ ExecInitSubPlan(SubPlan *subplan, PlanState *parent)
 	sstate->tab_eq_funcoids = NULL;
 	sstate->tab_hash_funcs = NULL;
 	sstate->tab_collations = NULL;
-	sstate->lhs_hash_funcs = NULL;
 	sstate->cur_eq_funcs = NULL;
 
 	/*
@@ -897,6 +896,7 @@ ExecInitSubPlan(SubPlan *subplan, PlanState *parent)
 		TupleDesc	tupDescRight;
 		Oid		   *cross_eq_funcoids;
 		TupleTableSlot *slot;
+		FmgrInfo   *lhs_hash_funcs;
 		List	   *oplist,
 				   *lefttlist,
 				   *righttlist;
@@ -953,7 +953,7 @@ ExecInitSubPlan(SubPlan *subplan, PlanState *parent)
 		sstate->tab_eq_funcoids = (Oid *) palloc(ncols * sizeof(Oid));
 		sstate->tab_collations = (Oid *) palloc(ncols * sizeof(Oid));
 		sstate->tab_hash_funcs = (FmgrInfo *) palloc(ncols * sizeof(FmgrInfo));
-		sstate->lhs_hash_funcs = (FmgrInfo *) palloc(ncols * sizeof(FmgrInfo));
+		lhs_hash_funcs = (FmgrInfo *) palloc(ncols * sizeof(FmgrInfo));
 		sstate->cur_eq_funcs = (FmgrInfo *) palloc(ncols * sizeof(FmgrInfo));
 		/* we'll need the cross-type equality fns below, but not in sstate */
 		cross_eq_funcoids = (Oid *) palloc(ncols * sizeof(Oid));
@@ -1003,7 +1003,7 @@ ExecInitSubPlan(SubPlan *subplan, PlanState *parent)
 									   &left_hashfn, &right_hashfn))
 				elog(ERROR, "could not find hash function for hash operator %u",
 					 opexpr->opno);
-			fmgr_info(left_hashfn, &sstate->lhs_hash_funcs[i - 1]);
+			fmgr_info(left_hashfn, &lhs_hash_funcs[i - 1]);
 			fmgr_info(right_hashfn, &sstate->tab_hash_funcs[i - 1]);
 
 			/* Set collation */
@@ -1039,6 +1039,16 @@ ExecInitSubPlan(SubPlan *subplan, PlanState *parent)
 													sstate->planstate,
 													NULL);
 
+		/* Build the ExprState for generating hash values */
+		sstate->lhs_hash_expr = ExecBuildHash32FromAttrs(tupDescLeft,
+														 &TTSOpsVirtual,
+														 lhs_hash_funcs,
+														 sstate->tab_collations,
+														 sstate->numCols,
+														 sstate->keyColIdx,
+														 parent,
+														 0);
+
 		/*
 		 * Create comparator for lookups of rows in the table (potentially
 		 * cross-type comparisons).