Fix planner to restore its previous level of intelligence about pushing

constants through full joins, as in

	select * from tenk1 a full join tenk1 b using (unique1)
	where unique1 = 42;

which should generate a fairly cheap plan where we apply the constraint
unique1 = 42 in each relation scan.  This had been broken by my patch of
2008-06-27, which is now reverted in favor of a more invasive but hopefully
less incorrect approach.  That patch was meant to prevent incorrect extraction
of OR'd indexclauses from OR conditions above an outer join.  To do that
correctly we need more information than the outerjoin_delay flag can provide,
so add a nullable_relids field to RestrictInfo that records exactly which
relations are nulled by outer joins that are underneath a particular qual
clause.  A side benefit is that we can make the test in create_or_index_quals
more specific: it is now smart enough to extract an OR'd indexclause into the
outer side of an outer join, even though it must not do so in the inner side.
The old coding couldn't distinguish these cases so it could not do either.

Author: tglsfdc

src/backend/nodes/copyfuncs.c

@@ -15,7 +15,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/nodes/copyfuncs.c,v 1.429 2009/04/05 19:59:39 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/nodes/copyfuncs.c,v 1.430 2009/04/16 20:42:16 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1606,6 +1606,7 @@ _copyRestrictInfo(RestrictInfo *from)
 	COPY_SCALAR_FIELD(pseudoconstant);
 	COPY_BITMAPSET_FIELD(clause_relids);
 	COPY_BITMAPSET_FIELD(required_relids);
+	COPY_BITMAPSET_FIELD(nullable_relids);
 	COPY_BITMAPSET_FIELD(left_relids);
 	COPY_BITMAPSET_FIELD(right_relids);
 	COPY_NODE_FIELD(orclause);

src/backend/nodes/equalfuncs.c

@@ -22,7 +22,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/nodes/equalfuncs.c,v 1.352 2009/04/05 19:59:40 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/nodes/equalfuncs.c,v 1.353 2009/04/16 20:42:16 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -759,6 +759,7 @@ _equalRestrictInfo(RestrictInfo *a, RestrictInfo *b)
 	COMPARE_SCALAR_FIELD(is_pushed_down);
 	COMPARE_SCALAR_FIELD(outerjoin_delayed);
 	COMPARE_BITMAPSET_FIELD(required_relids);
+	COMPARE_BITMAPSET_FIELD(nullable_relids);
 
 	/*
 	 * We ignore all the remaining fields, since they may not be set yet, and

src/backend/nodes/outfuncs.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.358 2009/04/05 19:59:40 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.359 2009/04/16 20:42:16 tgl Exp $
  *
  * NOTES
  *	  Every node type that can appear in stored rules' parsetrees *must*
@@ -1613,6 +1613,7 @@ _outRestrictInfo(StringInfo str, RestrictInfo *node)
 	WRITE_BOOL_FIELD(pseudoconstant);
 	WRITE_BITMAPSET_FIELD(clause_relids);
 	WRITE_BITMAPSET_FIELD(required_relids);
+	WRITE_BITMAPSET_FIELD(nullable_relids);
 	WRITE_BITMAPSET_FIELD(left_relids);
 	WRITE_BITMAPSET_FIELD(right_relids);
 	WRITE_NODE_FIELD(orclause);

src/backend/optimizer/path/indxpath.c

@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/path/indxpath.c,v 1.238 2009/03/11 03:32:22 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/path/indxpath.c,v 1.239 2009/04/16 20:42:16 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -2325,11 +2325,7 @@ expand_indexqual_conditions(IndexOptInfo *index, List *clausegroups)
 				if (boolqual)
 				{
 					resultquals = lappend(resultquals,
-										  make_restrictinfo(boolqual,
-															true,
-															false,
-															false,
-															NULL));
+										  make_simple_restrictinfo(boolqual));
 					continue;
 				}
 			}
@@ -2360,11 +2356,7 @@ expand_indexqual_conditions(IndexOptInfo *index, List *clausegroups)
 			{
 				Assert(index->amsearchnulls);
 				resultquals = lappend(resultquals,
-									  make_restrictinfo(clause,
-														true,
-														false,
-														false,
-														NULL));
+									  make_simple_restrictinfo(clause));
 			}
 			else
 				elog(ERROR, "unsupported indexqual type: %d",
@@ -2737,7 +2729,7 @@ expand_indexqual_rowcompare(RestrictInfo *rinfo,
 								  matching_cols);
 		rc->rargs = list_truncate((List *) copyObject(clause->rargs),
 								  matching_cols);
-		return make_restrictinfo((Expr *) rc, true, false, false, NULL);
+		return make_simple_restrictinfo((Expr *) rc);
 	}
 	else
 	{
@@ -2746,7 +2738,7 @@ expand_indexqual_rowcompare(RestrictInfo *rinfo,
 		opexpr = make_opclause(linitial_oid(new_ops), BOOLOID, false,
 							   copyObject(linitial(clause->largs)),
 							   copyObject(linitial(clause->rargs)));
-		return make_restrictinfo(opexpr, true, false, false, NULL);
+		return make_simple_restrictinfo(opexpr);
 	}
 }
 
@@ -2832,7 +2824,7 @@ prefix_quals(Node *leftop, Oid opfamily,
 			elog(ERROR, "no = operator for opfamily %u", opfamily);
 		expr = make_opclause(oproid, BOOLOID, false,
 							 (Expr *) leftop, (Expr *) prefix_const);
-		result = list_make1(make_restrictinfo(expr, true, false, false, NULL));
+		result = list_make1(make_simple_restrictinfo(expr));
 		return result;
 	}
 
@@ -2847,7 +2839,7 @@ prefix_quals(Node *leftop, Oid opfamily,
 		elog(ERROR, "no >= operator for opfamily %u", opfamily);
 	expr = make_opclause(oproid, BOOLOID, false,
 						 (Expr *) leftop, (Expr *) prefix_const);
-	result = list_make1(make_restrictinfo(expr, true, false, false, NULL));
+	result = list_make1(make_simple_restrictinfo(expr));
 
 	/*-------
 	 * If we can create a string larger than the prefix, we can say
@@ -2864,8 +2856,7 @@ prefix_quals(Node *leftop, Oid opfamily,
 	{
 		expr = make_opclause(oproid, BOOLOID, false,
 							 (Expr *) leftop, (Expr *) greaterstr);
-		result = lappend(result,
-						 make_restrictinfo(expr, true, false, false, NULL));
+		result = lappend(result, make_simple_restrictinfo(expr));
 	}
 
 	return result;
@@ -2928,7 +2919,7 @@ network_prefix_quals(Node *leftop, Oid expr_op, Oid opfamily, Datum rightop)
 						 (Expr *) leftop,
 						 (Expr *) makeConst(datatype, -1, -1, opr1right,
 											false, false));
-	result = list_make1(make_restrictinfo(expr, true, false, false, NULL));
+	result = list_make1(make_simple_restrictinfo(expr));
 
 	/* create clause "key <= network_scan_last( rightop )" */
 
@@ -2943,8 +2934,7 @@ network_prefix_quals(Node *leftop, Oid expr_op, Oid opfamily, Datum rightop)
 						 (Expr *) leftop,
 						 (Expr *) makeConst(datatype, -1, -1, opr2right,
 											false, false));
-	result = lappend(result,
-					 make_restrictinfo(expr, true, false, false, NULL));
+	result = lappend(result, make_simple_restrictinfo(expr));
 
 	return result;
 }

src/backend/optimizer/path/orindxpath.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/path/orindxpath.c,v 1.88 2009/02/15 20:16:21 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/path/orindxpath.c,v 1.89 2009/04/16 20:42:16 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -89,20 +89,26 @@ create_or_index_quals(PlannerInfo *root, RelOptInfo *rel)
 	ListCell   *i;
 
 	/*
-	 * Find potentially interesting OR joinclauses.  Note we must ignore any
-	 * joinclauses that are marked outerjoin_delayed or !is_pushed_down,
-	 * because they cannot be pushed down to the per-relation level due to
-	 * outer-join rules.  (XXX in some cases it might be possible to allow
-	 * this, but it would require substantially more bookkeeping about where
-	 * the clause came from.)
+	 * Find potentially interesting OR joinclauses.
+	 *
+	 * We must ignore clauses for which the target rel is in nullable_relids;
+	 * that means there's an outer join below the clause and so it can't be
+	 * enforced at the relation scan level.
+	 *
+	 * We must also ignore clauses that are marked !is_pushed_down (ie they
+	 * are themselves outer-join clauses).  It would be safe to extract an
+	 * index condition from such a clause if we are within the nullable rather
+	 * than the non-nullable side of its join, but we haven't got enough
+	 * context here to tell which applies.  OR clauses in outer-join quals
+	 * aren't exactly common, so we'll let that case go unoptimized for now.
 	 */
 	foreach(i, rel->joininfo)
 	{
 		RestrictInfo *rinfo = (RestrictInfo *) lfirst(i);
 
 		if (restriction_is_or_clause(rinfo) &&
 			rinfo->is_pushed_down &&
-			!rinfo->outerjoin_delayed)
+			!bms_is_member(rel->relid, rinfo->nullable_relids))
 		{
 			/*
 			 * Use the generate_bitmap_or_paths() machinery to estimate the

src/backend/optimizer/plan/initsplan.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/plan/initsplan.c,v 1.149 2009/02/27 22:41:38 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/plan/initsplan.c,v 1.150 2009/04/16 20:42:16 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -53,7 +53,7 @@ static void distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 						Relids ojscope,
 						Relids outerjoin_nonnullable);
 static bool check_outerjoin_delay(PlannerInfo *root, Relids *relids_p,
-					  bool is_pushed_down);
+					  Relids *nullable_relids_p, bool is_pushed_down);
 static bool check_redundant_nullability_qual(PlannerInfo *root, Node *clause);
 static void check_mergejoinable(RestrictInfo *restrictinfo);
 static void check_hashjoinable(RestrictInfo *restrictinfo);
@@ -755,6 +755,7 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 	bool		pseudoconstant = false;
 	bool		maybe_equivalence;
 	bool		maybe_outer_join;
+	Relids		nullable_relids;
 	RestrictInfo *restrictinfo;
 
 	/*
@@ -861,6 +862,7 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 		Assert(!ojscope);
 		is_pushed_down = true;
 		outerjoin_delayed = false;
+		nullable_relids = NULL;
 		/* Don't feed it back for more deductions */
 		maybe_equivalence = false;
 		maybe_outer_join = false;
@@ -882,7 +884,10 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 		maybe_outer_join = true;
 
 		/* Check to see if must be delayed by lower outer join */
-		outerjoin_delayed = check_outerjoin_delay(root, &relids, false);
+		outerjoin_delayed = check_outerjoin_delay(root,
+												  &relids,
+												  &nullable_relids,
+												  false);
 
 		/*
 		 * Now force the qual to be evaluated exactly at the level of joining
@@ -907,7 +912,10 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 		is_pushed_down = true;
 
 		/* Check to see if must be delayed by lower outer join */
-		outerjoin_delayed = check_outerjoin_delay(root, &relids, true);
+		outerjoin_delayed = check_outerjoin_delay(root,
+												  &relids,
+												  &nullable_relids,
+												  true);
 
 		if (outerjoin_delayed)
 		{
@@ -957,7 +965,8 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
 									 is_pushed_down,
 									 outerjoin_delayed,
 									 pseudoconstant,
-									 relids);
+									 relids,
+									 nullable_relids);
 
 	/*
 	 * If it's a join clause (either naturally, or because delayed by
@@ -1064,7 +1073,9 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
  * If the qual must be delayed, add relids to *relids_p to reflect the lowest
  * safe level for evaluating the qual, and return TRUE.  Any extra delay for
  * higher-level joins is reflected by setting delay_upper_joins to TRUE in
- * SpecialJoinInfo structs.
+ * SpecialJoinInfo structs.  We also compute nullable_relids, the set of
+ * referenced relids that are nullable by lower outer joins (note that this
+ * can be nonempty even for a non-delayed qual).
  *
  * For an is_pushed_down qual, we can evaluate the qual as soon as (1) we have
  * all the rels it mentions, and (2) we are at or above any outer joins that
@@ -1087,8 +1098,8 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
  * mentioning only C cannot be applied below the join to A.
  *
  * For a non-pushed-down qual, this isn't going to determine where we place the
- * qual, but we need to determine outerjoin_delayed anyway for possible use
- * in reconsider_outer_join_clauses().
+ * qual, but we need to determine outerjoin_delayed and nullable_relids anyway
+ * for use later in the planning process.
  *
  * Lastly, a pushed-down qual that references the nullable side of any current
  * join_info_list member and has to be evaluated above that OJ (because its
@@ -1104,13 +1115,26 @@ distribute_qual_to_rels(PlannerInfo *root, Node *clause,
  * two OJs to commute.)
  */
 static bool
-check_outerjoin_delay(PlannerInfo *root, Relids *relids_p,
+check_outerjoin_delay(PlannerInfo *root,
+					  Relids *relids_p,				/* in/out parameter */
+					  Relids *nullable_relids_p,	/* output parameter */
 					  bool is_pushed_down)
 {
-	Relids		relids = *relids_p;
+	Relids		relids;
+	Relids		nullable_relids;
 	bool		outerjoin_delayed;
 	bool		found_some;
 
+	/* fast path if no special joins */
+	if (root->join_info_list == NIL)
+	{
+		*nullable_relids_p = NULL;
+		return false;
+	}
+
+	/* must copy relids because we need the original value at the end */
+	relids = bms_copy(*relids_p);
+	nullable_relids = NULL;
 	outerjoin_delayed = false;
 	do
 	{
@@ -1126,18 +1150,23 @@ check_outerjoin_delay(PlannerInfo *root, Relids *relids_p,
 				(sjinfo->jointype == JOIN_FULL &&
 				 bms_overlap(relids, sjinfo->min_lefthand)))
 			{
-				/* yes, so set the result flag */
-				outerjoin_delayed = true;
-				/* have we included all its rels in relids? */
+				/* yes; have we included all its rels in relids? */
 				if (!bms_is_subset(sjinfo->min_lefthand, relids) ||
 					!bms_is_subset(sjinfo->min_righthand, relids))
 				{
 					/* no, so add them in */
 					relids = bms_add_members(relids, sjinfo->min_lefthand);
 					relids = bms_add_members(relids, sjinfo->min_righthand);
+					outerjoin_delayed = true;
 					/* we'll need another iteration */
 					found_some = true;
 				}
+				/* track all the nullable rels of relevant OJs */
+				nullable_relids = bms_add_members(nullable_relids,
+												  sjinfo->min_righthand);
+				if (sjinfo->jointype == JOIN_FULL)
+					nullable_relids = bms_add_members(nullable_relids,
+													  sjinfo->min_lefthand);
 				/* set delay_upper_joins if needed */
 				if (is_pushed_down && sjinfo->jointype != JOIN_FULL &&
 					bms_overlap(relids, sjinfo->min_lefthand))
@@ -1146,7 +1175,13 @@ check_outerjoin_delay(PlannerInfo *root, Relids *relids_p,
 		}
 	} while (found_some);
 
+	/* identify just the actually-referenced nullable rels */
+	nullable_relids = bms_int_members(nullable_relids, *relids_p);
+
+	/* replace *relids_p, and return nullable_relids */
+	bms_free(*relids_p);
 	*relids_p = relids;
+	*nullable_relids_p = nullable_relids;
 	return outerjoin_delayed;
 }
 
@@ -1352,7 +1387,8 @@ build_implied_join_equality(Oid opno,
 									 true,		/* is_pushed_down */
 									 false,		/* outerjoin_delayed */
 									 false,		/* pseudoconstant */
-									 qualscope);
+									 qualscope,	/* required_relids */
+									 NULL);		/* nullable_relids */
 
 	/* Set mergejoinability info always, and hashjoinability if enabled */
 	check_mergejoinable(restrictinfo);