Compute correct em_nullable_relids in get_eclass_for_sort_expr().

Bug #8591 from Claudio Freire demonstrates that get_eclass_for_sort_expr
must be able to compute valid em_nullable_relids for any new equivalence
class members it creates.  I'd worried about this in the commit message
for db9f0e1d9a4a0842c814a464cdc9758c3f20b96c, but claimed that it wasn't a
problem because multi-member ECs should already exist when it runs.  That
is transparently wrong, though, because this function is also called by
initialize_mergeclause_eclasses, which runs during deconstruct_jointree.
The example given in the bug report (which the new regression test item
is based upon) fails because the COALESCE() expression is first seen by
initialize_mergeclause_eclasses rather than process_equivalence.

Fixing this requires passing the appropriate nullable_relids set to
get_eclass_for_sort_expr, and it requires new code to compute that set
for top-level expressions such as ORDER BY, GROUP BY, etc.  We store
the top-level nullable_relids in a new field in PlannerInfo to avoid
computing it many times.  In the back branches, I've added the new
field at the end of the struct to minimize ABI breakage for planner
plugins.  There doesn't seem to be a good alternative to changing
get_eclass_for_sort_expr's API signature, though.  There probably aren't
any third-party extensions calling that function directly; moreover,
if there are, they probably need to think about what to pass for
nullable_relids anyway.

Back-patch to 9.2, like the previous patch in this area.

1 files changed, 20 insertions, 0 deletions

diff --git a/src/backend/optimizer/plan/initsplan.c b/src/backend/optimizer/plan/initsplan.c
index c5998b9e2de..04a399ee13c 100644
--- a/src/backend/optimizer/plan/initsplan.c
+++ b/src/backend/optimizer/plan/initsplan.c
@@ -649,6 +649,9 @@ deconstruct_jointree(PlannerInfo *root)
 	Assert(root->parse->jointree != NULL &&
 		   IsA(root->parse->jointree, FromExpr));
 
+	/* this is filled as we scan the jointree */
+	root->nullable_baserels = NULL;
+
 	result = deconstruct_recurse(root, (Node *) root->parse->jointree, false,
 								 &qualscope, &inner_join_rels,
 								 &postponed_qual_list);
@@ -790,6 +793,7 @@ deconstruct_recurse(PlannerInfo *root, Node *jtnode, bool below_outer_join,
 					left_inners,
 					right_inners,
 					nonnullable_rels,
+					nullable_rels,
 					ojscope;
 		List	   *leftjoinlist,
 				   *rightjoinlist;
@@ -823,6 +827,8 @@ deconstruct_recurse(PlannerInfo *root, Node *jtnode, bool below_outer_join,
 				*inner_join_rels = *qualscope;
 				/* Inner join adds no restrictions for quals */
 				nonnullable_rels = NULL;
+				/* and it doesn't force anything to null, either */
+				nullable_rels = NULL;
 				break;
 			case JOIN_LEFT:
 			case JOIN_ANTI:
@@ -837,6 +843,7 @@ deconstruct_recurse(PlannerInfo *root, Node *jtnode, bool below_outer_join,
 				*qualscope = bms_union(leftids, rightids);
 				*inner_join_rels = bms_union(left_inners, right_inners);
 				nonnullable_rels = leftids;
+				nullable_rels = rightids;
 				break;
 			case JOIN_SEMI:
 				leftjoinlist = deconstruct_recurse(root, j->larg,
@@ -851,6 +858,13 @@ deconstruct_recurse(PlannerInfo *root, Node *jtnode, bool below_outer_join,
 				*inner_join_rels = bms_union(left_inners, right_inners);
 				/* Semi join adds no restrictions for quals */
 				nonnullable_rels = NULL;
+
+				/*
+				 * Theoretically, a semijoin would null the RHS; but since the
+				 * RHS can't be accessed above the join, this is immaterial
+				 * and we needn't account for it.
+				 */
+				nullable_rels = NULL;
 				break;
 			case JOIN_FULL:
 				leftjoinlist = deconstruct_recurse(root, j->larg,
@@ -865,16 +879,22 @@ deconstruct_recurse(PlannerInfo *root, Node *jtnode, bool below_outer_join,
 				*inner_join_rels = bms_union(left_inners, right_inners);
 				/* each side is both outer and inner */
 				nonnullable_rels = *qualscope;
+				nullable_rels = *qualscope;
 				break;
 			default:
 				/* JOIN_RIGHT was eliminated during reduce_outer_joins() */
 				elog(ERROR, "unrecognized join type: %d",
 					 (int) j->jointype);
 				nonnullable_rels = NULL;		/* keep compiler quiet */
+				nullable_rels = NULL;
 				leftjoinlist = rightjoinlist = NIL;
 				break;
 		}
 
+		/* Report all rels that will be nulled anywhere in the jointree */
+		root->nullable_baserels = bms_add_members(root->nullable_baserels,
+												  nullable_rels);
+
 		/*
 		 * For an OJ, form the SpecialJoinInfo now, because we need the OJ's
 		 * semantic scope (ojscope) to pass to distribute_qual_to_rels.  But