If we expect a hash join to be performed in multiple batches, suppress

"physical tlist" optimization on the outer relation (ie, force a projection
step to occur in its scan).  This avoids storing useless column values when
the outer relation's tuples are written to temporary batch files.

Modified version of a patch by Michael Henderson and Ramon Lawrence.

Author: tglsfdc

src/backend/nodes/outfuncs.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.355 2009/03/21 00:04:39 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/nodes/outfuncs.c,v 1.356 2009/03/26 17:15:34 tgl Exp $
  *
  * NOTES
  *	  Every node type that can appear in stored rules' parsetrees *must*
@@ -1448,6 +1448,7 @@ _outHashPath(StringInfo str, HashPath *node)
 	_outJoinPathInfo(str, (JoinPath *) node);
 
 	WRITE_NODE_FIELD(path_hashclauses);
+	WRITE_INT_FIELD(num_batches);
 }
 
 static void

src/backend/optimizer/path/costsize.c

@@ -54,7 +54,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.205 2009/03/21 00:04:39 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/path/costsize.c,v 1.206 2009/03/26 17:15:35 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1880,6 +1880,8 @@ cost_hashjoin(HashPath *path, PlannerInfo *root, SpecialJoinInfo *sjinfo)
 							&numbatches,
 							&num_skew_mcvs);
 	virtualbuckets = (double) numbuckets *(double) numbatches;
+	/* mark the path with estimated # of batches */
+	path->num_batches = numbatches;
 
 	/*
 	 * Determine bucketsize fraction for inner relation.  We use the smallest

src/backend/optimizer/plan/createplan.c

@@ -10,7 +10,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.256 2009/03/21 00:04:39 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/plan/createplan.c,v 1.257 2009/03/26 17:15:35 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1910,6 +1910,10 @@ create_hashjoin_plan(PlannerInfo *root,
 	/* We don't want any excess columns in the hashed tuples */
 	disuse_physical_tlist(inner_plan, best_path->jpath.innerjoinpath);
 
+	/* If we expect batching, suppress excess columns in outer tuples too */
+	if (best_path->num_batches > 1)
+		disuse_physical_tlist(outer_plan, best_path->jpath.outerjoinpath);
+
 	/*
 	 * If there is a single join clause and we can identify the outer
 	 * variable as a simple column reference, supply its identity for

src/backend/optimizer/util/pathnode.c

@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *	  $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.150 2009/02/27 00:06:27 tgl Exp $
+ *	  $PostgreSQL: pgsql/src/backend/optimizer/util/pathnode.c,v 1.151 2009/03/26 17:15:35 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -1480,9 +1480,20 @@ create_hashjoin_path(PlannerInfo *root,
 	pathnode->jpath.outerjoinpath = outer_path;
 	pathnode->jpath.innerjoinpath = inner_path;
 	pathnode->jpath.joinrestrictinfo = restrict_clauses;
-	/* A hashjoin never has pathkeys, since its ordering is unpredictable */
+	/*
+	 * A hashjoin never has pathkeys, since its output ordering is
+	 * unpredictable due to possible batching.  XXX If the inner relation is
+	 * small enough, we could instruct the executor that it must not batch,
+	 * and then we could assume that the output inherits the outer relation's
+	 * ordering, which might save a sort step.  However there is considerable
+	 * downside if our estimate of the inner relation size is badly off.
+	 * For the moment we don't risk it.  (Note also that if we wanted to take
+	 * this seriously, joinpath.c would have to consider many more paths for
+	 * the outer rel than it does now.)
+	 */
 	pathnode->jpath.path.pathkeys = NIL;
 	pathnode->path_hashclauses = hashclauses;
+	/* cost_hashjoin will fill in pathnode->num_batches */
 
 	cost_hashjoin(pathnode, root, sjinfo);
 

src/include/nodes/relation.h

@@ -7,7 +7,7 @@
  * Portions Copyright (c) 1996-2009, PostgreSQL Global Development Group
  * Portions Copyright (c) 1994, Regents of the University of California
  *
- * $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.170 2009/03/05 23:06:45 tgl Exp $
+ * $PostgreSQL: pgsql/src/include/nodes/relation.h,v 1.171 2009/03/26 17:15:35 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -845,6 +845,7 @@ typedef struct HashPath
 {
 	JoinPath	jpath;
 	List	   *path_hashclauses;		/* join clauses used for hashing */
+	int			num_batches;			/* number of batches expected */
 } HashPath;
 
 /*