Fix thinko in new match_join_clauses_to_index() logic.

tglsfdc · tglsfdc · commit 0140a11b9ba5 · 2012-02-28T18:10:40.000-05:00
We don't need to constrain the other side of an indexable join clause to
not be below an outer join; an example here is

SELECT FROM t1 LEFT JOIN t2 ON t1.a = t2.b LEFT JOIN t3 ON t2.c = t3.d;

We can consider an inner indexscan on t3.d using c = d as indexqual, even
though t2.c is potentially nulled by a previous outer join.  The comparable
logic in orindxpath.c has always worked that way, but I was being overly
cautious here.
diff --git a/src/backend/optimizer/path/indxpath.c b/src/backend/optimizer/path/indxpath.c
@@ -1702,17 +1702,17 @@ match_join_clauses_to_index(PlannerInfo *root,
 		 * outer join rules.
 		 *
 		 * Instead of considering required_relids, we ignore clauses for which
-		 * any referenced rel is in nullable_relids; that means there's an
-		 * outer join below the clause and so it can't be checked at the
-		 * relation scan level.
+		 * the indexed rel is in nullable_relids; that means there's an outer
+		 * join below the clause and so it can't be checked at the relation
+		 * scan level.
 		 *
 		 * Note: unlike create_or_index_quals(), we can accept clauses that
 		 * are marked !is_pushed_down (ie they are themselves outer-join
 		 * clauses).  This is OK because any path generated with these clauses
 		 * could only be used in the inside of a nestloop join, which will be
 		 * the nullable side.
 		 */
-		if (bms_overlap(rinfo->clause_relids, rinfo->nullable_relids))
+		if (bms_overlap(rel->relids, rinfo->nullable_relids))
 			continue;
 
 		/* Potentially usable, so see if it matches the index or is an OR */
