Fix hypertables not getting removed from useless outer joins

svenklemm · svenklemm · commit c300d2eee6f5 · 2026-07-21T20:48:18.000+02:00
If a hypertable only appears on the nullable side of a LEFT/RIGHT/FULL join, and none of its columns are used elsewhere in the query, Postgres can normally drop the join entirely. This stopped working after 390d900 (Run our hypertable expansion only from the get_relation_info_hook, #9714), which removed marking a hypertable early enough for Postgres to still see it as a plain table.
diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
@@ -136,6 +136,7 @@ set(IMPORTED_SOURCES
     import/heapswap.c
     import/list.c
     import/planner.c
+    import/prepjointree.c
     import/setrefs.c
     import/ts_explain.c
     import/ts_inherit.c)
diff --git a/src/import/prepjointree.c b/src/import/prepjointree.c
@@ -0,0 +1,141 @@
+/*
+ * This file and its contents are licensed under the Apache License 2.0.
+ * Please see the included NOTICE for copyright information and
+ * LICENSE-APACHE for a copy of the license.
+ */
+
+/*
+ * This file contains source code that was copied and/or modified from
+ * the PostgreSQL database, which is licensed under the open-source
+ * PostgreSQL License. Please see the NOTICE at the top level
+ * directory for a copy of the PostgreSQL License.
+ */
+#include <postgres.h>
+#include <nodes/bitmapset.h>
+#include <nodes/nodes.h>
+#include <nodes/parsenodes.h>
+#include <nodes/pathnodes.h>
+#include <nodes/primnodes.h>
+
+#include "prepjointree.h"
+
+/*
+ * Copied verbatim from src/backend/optimizer/prep/prepjointree.c.
+ */
+typedef struct nullingrel_info
+{
+	/*
+	 * For each leaf RTE, nullingrels[rti] is the set of relids of outer joins
+	 * that potentially null that RTE.
+	 */
+	Relids *nullingrels;
+	/* Length of range table (maximum index in nullingrels[]) */
+	int rtlength; /* used only for assertion checks */
+} nullingrel_info;
+
+static void get_nullingrels_recurse(Node *jtnode, Relids upper_nullingrels,
+									nullingrel_info *info);
+
+/*
+ * get_nullingrels: collect info about which outer joins null which relations
+ *
+ * The result struct contains, for each leaf relation used in the query,
+ * the set of relids of outer joins that potentially null that rel.
+ */
+static nullingrel_info *
+get_nullingrels(Query *parse)
+{
+	nullingrel_info *result = palloc_object(nullingrel_info);
+
+	result->rtlength = list_length(parse->rtable);
+	result->nullingrels = palloc0_array(Relids, result->rtlength + 1);
+	get_nullingrels_recurse((Node *) parse->jointree, NULL, result);
+	return result;
+}
+
+/*
+ * Recursive guts of get_nullingrels().
+ *
+ * Note: at any recursion level, the passed-down upper_nullingrels must be
+ * treated as a constant, but it can be stored directly into *info
+ * if we're at leaf level.  Upper recursion levels do not free their mutated
+ * copies of the nullingrels, because those are probably referenced by
+ * at least one leaf rel.
+ */
+static void
+get_nullingrels_recurse(Node *jtnode, Relids upper_nullingrels, nullingrel_info *info)
+{
+	if (jtnode == NULL)
+		return;
+	if (IsA(jtnode, RangeTblRef))
+	{
+		int varno = ((RangeTblRef *) jtnode)->rtindex;
+
+		Assert(varno > 0 && varno <= info->rtlength);
+		info->nullingrels[varno] = upper_nullingrels;
+	}
+	else if (IsA(jtnode, FromExpr))
+	{
+		FromExpr *f = (FromExpr *) jtnode;
+		ListCell *l;
+
+		foreach (l, f->fromlist)
+		{
+			get_nullingrels_recurse(lfirst(l), upper_nullingrels, info);
+		}
+	}
+	else if (IsA(jtnode, JoinExpr))
+	{
+		JoinExpr *j = (JoinExpr *) jtnode;
+		Relids local_nullingrels;
+
+		switch (j->jointype)
+		{
+			case JOIN_INNER:
+				get_nullingrels_recurse(j->larg, upper_nullingrels, info);
+				get_nullingrels_recurse(j->rarg, upper_nullingrels, info);
+				break;
+			case JOIN_LEFT:
+			case JOIN_SEMI:
+			case JOIN_ANTI:
+				local_nullingrels = bms_add_member(bms_copy(upper_nullingrels), j->rtindex);
+				get_nullingrels_recurse(j->larg, upper_nullingrels, info);
+				get_nullingrels_recurse(j->rarg, local_nullingrels, info);
+				break;
+			case JOIN_FULL:
+				local_nullingrels = bms_add_member(bms_copy(upper_nullingrels), j->rtindex);
+				get_nullingrels_recurse(j->larg, local_nullingrels, info);
+				get_nullingrels_recurse(j->rarg, local_nullingrels, info);
+				break;
+			case JOIN_RIGHT:
+				local_nullingrels = bms_add_member(bms_copy(upper_nullingrels), j->rtindex);
+				get_nullingrels_recurse(j->larg, local_nullingrels, info);
+				get_nullingrels_recurse(j->rarg, upper_nullingrels, info);
+				break;
+			default:
+				elog(ERROR, "unrecognized join type: %d", (int) j->jointype);
+				break;
+		}
+	}
+	else
+		elog(ERROR, "unrecognized node type: %d", (int) nodeTag(jtnode));
+}
+
+/*
+ * Return the set of rtindexes of relations that can be NULL-extended by some
+ * outer join in the query's jointree.
+ */
+Bitmapset *
+ts_get_nullable_rtis(Query *parse)
+{
+	nullingrel_info *info = get_nullingrels(parse);
+	Bitmapset *nullable_rtis = NULL;
+
+	for (int rti = 1; rti <= info->rtlength; rti++)
+	{
+		if (!bms_is_empty(info->nullingrels[rti]))
+			nullable_rtis = bms_add_member(nullable_rtis, rti);
+	}
+
+	return nullable_rtis;
+}
diff --git a/src/import/prepjointree.h b/src/import/prepjointree.h
@@ -0,0 +1,19 @@
+/*
+ * This file and its contents are licensed under the Apache License 2.0.
+ * Please see the included NOTICE for copyright information and
+ * LICENSE-APACHE for a copy of the license.
+ */
+
+/*
+ * This file contains source code that was copied and/or modified from
+ * the PostgreSQL database, which is licensed under the open-source
+ * PostgreSQL License. Please see the NOTICE at the top level
+ * directory for a copy of the PostgreSQL License.
+ */
+#pragma once
+
+#include <postgres.h>
+#include <nodes/bitmapset.h>
+#include <nodes/parsenodes.h>
+
+extern Bitmapset *ts_get_nullable_rtis(Query *parse);
diff --git a/src/planner/planner.c b/src/planner/planner.c
@@ -50,6 +50,7 @@
 #include "hypertable.h"
 #include "hypertable_cache.h"
 #include "import/allpaths.h"
+#include "import/prepjointree.h"
 #include "license_guc.h"
 #include "nodes/chunk_append/chunk_append.h"
 #include "nodes/constraint_aware_append/constraint_aware_append.h"
@@ -444,6 +445,7 @@ preprocess_query(Node *node, PreprocessQueryContext *context)
 		ListCell *lc;
 		Index rti = 1;
 		bool ret;
+		Bitmapset *nullable_rtis = ts_get_nullable_rtis(query);
 
 		if (ts_guc_enable_foreign_key_propagation)
 		{
@@ -490,6 +492,21 @@ preprocess_query(Node *node, PreprocessQueryContext *context)
 							rte_mark_for_expansion(rte);
 						}
 					}
+					else if (bms_is_member(rti, nullable_rtis))
+					{
+						/*
+						 * Mark hypertable RTEs on the nullable side of an outer
+						 * join here too: get_relation_info_hook runs after join
+						 * removal, so a hypertable only visible via view inlining
+						 * would still look like an inheritance parent when join
+						 * removal runs, blocking useless-join elimination.
+						 */
+						if (ts_guc_enable_optimizations && ts_guc_enable_constraint_exclusion &&
+							rte->inh && (Index) query->resultRelation != rti)
+						{
+							rte_mark_for_expansion(rte);
+						}
+					}
 					break;
 				default:
 					break;
diff --git a/test/expected/plan_expand_hypertable-16.out b/test/expected/plan_expand_hypertable-16.out
@@ -3671,4 +3671,26 @@ EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT * FROM
    ->  Index Scan using _hyper_21_203_chunk_metrics_space_time_idx on _hyper_21_203_chunk (actual rows=1.00 loops=1)
          Index Cond: ("time" >= 'Sun Jan 02 00:00:00 2000 UTC'::timestamp with time zone)
 
+CREATE TABLE join_removal_main(time timestamptz NOT NULL, device int, val float);
+SELECT create_hypertable('join_removal_main', 'time');
+        create_hypertable        
+---------------------------------
+ (22,public,join_removal_main,t)
+
+CREATE TABLE join_removal_aux(time timestamptz NOT NULL, device int NOT NULL, extra float);
+SELECT create_hypertable('join_removal_aux', 'time');
+       create_hypertable        
+--------------------------------
+ (23,public,join_removal_aux,t)
+
+CREATE UNIQUE INDEX join_removal_aux_uq ON join_removal_aux (device, time);
+INSERT INTO join_removal_main VALUES ('2021-01-01', 1, 1.0), ('2021-01-02', 2, 2.0);
+INSERT INTO join_removal_aux VALUES ('2021-01-01', 1, 10.0), ('2021-01-02', 2, 20.0);
+ANALYZE join_removal_main, join_removal_aux;
+EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT m.time, m.device, m.val
+FROM join_removal_main m
+LEFT JOIN join_removal_aux a ON a.device = m.device AND a.time = m.time;
+--- QUERY PLAN ---
+ Seq Scan on _hyper_22_204_chunk m (actual rows=2.00 loops=1)
+
 --TEST END--
diff --git a/test/expected/plan_expand_hypertable-17.out b/test/expected/plan_expand_hypertable-17.out
@@ -3671,4 +3671,26 @@ EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT * FROM
    ->  Index Scan using _hyper_21_203_chunk_metrics_space_time_idx on _hyper_21_203_chunk (actual rows=1.00 loops=1)
          Index Cond: ("time" >= 'Sun Jan 02 00:00:00 2000 UTC'::timestamp with time zone)
 
+CREATE TABLE join_removal_main(time timestamptz NOT NULL, device int, val float);
+SELECT create_hypertable('join_removal_main', 'time');
+        create_hypertable        
+---------------------------------
+ (22,public,join_removal_main,t)
+
+CREATE TABLE join_removal_aux(time timestamptz NOT NULL, device int NOT NULL, extra float);
+SELECT create_hypertable('join_removal_aux', 'time');
+       create_hypertable        
+--------------------------------
+ (23,public,join_removal_aux,t)
+
+CREATE UNIQUE INDEX join_removal_aux_uq ON join_removal_aux (device, time);
+INSERT INTO join_removal_main VALUES ('2021-01-01', 1, 1.0), ('2021-01-02', 2, 2.0);
+INSERT INTO join_removal_aux VALUES ('2021-01-01', 1, 10.0), ('2021-01-02', 2, 20.0);
+ANALYZE join_removal_main, join_removal_aux;
+EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT m.time, m.device, m.val
+FROM join_removal_main m
+LEFT JOIN join_removal_aux a ON a.device = m.device AND a.time = m.time;
+--- QUERY PLAN ---
+ Seq Scan on _hyper_22_204_chunk m (actual rows=2.00 loops=1)
+
 --TEST END--
diff --git a/test/expected/plan_expand_hypertable-18.out b/test/expected/plan_expand_hypertable-18.out
@@ -3671,4 +3671,26 @@ EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT * FROM
    ->  Index Scan using _hyper_21_203_chunk_metrics_space_time_idx on _hyper_21_203_chunk (actual rows=1.00 loops=1)
          Index Cond: ("time" >= 'Sun Jan 02 00:00:00 2000 UTC'::timestamp with time zone)
 
+CREATE TABLE join_removal_main(time timestamptz NOT NULL, device int, val float);
+SELECT create_hypertable('join_removal_main', 'time');
+        create_hypertable        
+---------------------------------
+ (22,public,join_removal_main,t)
+
+CREATE TABLE join_removal_aux(time timestamptz NOT NULL, device int NOT NULL, extra float);
+SELECT create_hypertable('join_removal_aux', 'time');
+       create_hypertable        
+--------------------------------
+ (23,public,join_removal_aux,t)
+
+CREATE UNIQUE INDEX join_removal_aux_uq ON join_removal_aux (device, time);
+INSERT INTO join_removal_main VALUES ('2021-01-01', 1, 1.0), ('2021-01-02', 2, 2.0);
+INSERT INTO join_removal_aux VALUES ('2021-01-01', 1, 10.0), ('2021-01-02', 2, 20.0);
+ANALYZE join_removal_main, join_removal_aux;
+EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT m.time, m.device, m.val
+FROM join_removal_main m
+LEFT JOIN join_removal_aux a ON a.device = m.device AND a.time = m.time;
+--- QUERY PLAN ---
+ Seq Scan on _hyper_22_204_chunk m (actual rows=2.00 loops=1)
+
 --TEST END--
diff --git a/test/expected/plan_expand_hypertable-19.out b/test/expected/plan_expand_hypertable-19.out
@@ -3670,4 +3670,26 @@ EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT * FROM
    ->  Index Scan using _hyper_21_203_chunk_metrics_space_time_idx on _hyper_21_203_chunk (actual rows=1.00 loops=1)
          Index Cond: ("time" >= 'Sun Jan 02 00:00:00 2000 UTC'::timestamp with time zone)
 
+CREATE TABLE join_removal_main(time timestamptz NOT NULL, device int, val float);
+SELECT create_hypertable('join_removal_main', 'time');
+        create_hypertable        
+---------------------------------
+ (22,public,join_removal_main,t)
+
+CREATE TABLE join_removal_aux(time timestamptz NOT NULL, device int NOT NULL, extra float);
+SELECT create_hypertable('join_removal_aux', 'time');
+       create_hypertable        
+--------------------------------
+ (23,public,join_removal_aux,t)
+
+CREATE UNIQUE INDEX join_removal_aux_uq ON join_removal_aux (device, time);
+INSERT INTO join_removal_main VALUES ('2021-01-01', 1, 1.0), ('2021-01-02', 2, 2.0);
+INSERT INTO join_removal_aux VALUES ('2021-01-01', 1, 10.0), ('2021-01-02', 2, 20.0);
+ANALYZE join_removal_main, join_removal_aux;
+EXPLAIN (buffers off, costs off, timing off, summary off, analyze) SELECT m.time, m.device, m.val
+FROM join_removal_main m
+LEFT JOIN join_removal_aux a ON a.device = m.device AND a.time = m.time;
+--- QUERY PLAN ---
+ Seq Scan on _hyper_22_204_chunk m (actual rows=2.00 loops=1)
+
 --TEST END--
diff --git a/test/sql/plan_expand_hypertable.sql.in b/test/sql/plan_expand_hypertable.sql.in
@@ -172,4 +172,21 @@ INSERT INTO metrics_space SELECT '2000-01-01'::timestamptz + format('%s day',i):
 -- no contraints removed due to space partitioning
 :PREFIX SELECT * FROM metrics_space WHERE time >= '2000-01-02';
 
+-- test that a useless LEFT JOIN to a hypertable is eliminated when none of
+-- its columns are selected and the join clause matches a unique index
+CREATE TABLE join_removal_main(time timestamptz NOT NULL, device int, val float);
+SELECT create_hypertable('join_removal_main', 'time');
+
+CREATE TABLE join_removal_aux(time timestamptz NOT NULL, device int NOT NULL, extra float);
+SELECT create_hypertable('join_removal_aux', 'time');
+CREATE UNIQUE INDEX join_removal_aux_uq ON join_removal_aux (device, time);
+
+INSERT INTO join_removal_main VALUES ('2021-01-01', 1, 1.0), ('2021-01-02', 2, 2.0);
+INSERT INTO join_removal_aux VALUES ('2021-01-01', 1, 10.0), ('2021-01-02', 2, 20.0);
+ANALYZE join_removal_main, join_removal_aux;
+
+:PREFIX SELECT m.time, m.device, m.val
+FROM join_removal_main m
+LEFT JOIN join_removal_aux a ON a.device = m.device AND a.time = m.time;
+
 \qecho '--TEST END--'
