From b5ced403d86f1647529616f3108608fdb2362095 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Philip=20Dub=C3=A9?= <gunkmute@gmail.com>
Date: Fri, 31 May 2019 12:08:26 -0700
Subject: [PATCH] Also check rewrittenQuery jointree for outer join

---
 .../planner/query_pushdown_planning.c         | 13 +++++++++++++
 src/test/regress/expected/multi_subquery.out  | 19 +++++++++++++++++++
 .../expected/non_colocated_subquery_joins.out |  2 +-
 src/test/regress/sql/multi_subquery.sql       | 11 +++++++++++
 4 files changed, 44 insertions(+), 1 deletion(-)

diff --git a/src/backend/distributed/planner/query_pushdown_planning.c b/src/backend/distributed/planner/query_pushdown_planning.c
index a0cc7ace2..5aef6fde5 100644
--- a/src/backend/distributed/planner/query_pushdown_planning.c
+++ b/src/backend/distributed/planner/query_pushdown_planning.c
@@ -142,6 +142,19 @@ ShouldUseSubqueryPushDown(Query *originalQuery, Query *rewrittenQuery)
 		return true;
 	}
 
+	/*
+	 * Original query may not have an outer join while rewritten query does.
+	 * We should push down in this case.
+	 * An example of this is https://github.com/citusdata/citus/issues/2739
+	 * where postgres pulls-up the outer-join in the subquery.
+	 */
+	if (FindNodeCheck((Node *) rewrittenQuery->jointree, IsOuterJoinExpr))
+	{
+		/* Assert what _should_ be only situation this occurs in. */
+		Assert(JoinTreeContainsSubquery(originalQuery));
+		return true;
+	}
+
 	/*
 	 * Some unsupported join clauses in logical planner
 	 * may be supported by subquery pushdown planner.
diff --git a/src/test/regress/expected/multi_subquery.out b/src/test/regress/expected/multi_subquery.out
index e93cf0623..92ca1f226 100644
--- a/src/test/regress/expected/multi_subquery.out
+++ b/src/test/regress/expected/multi_subquery.out
@@ -462,6 +462,25 @@ SELECT DISTINCT ON (t1.user_id) t1.user_id, t2.value_1, t2.value_2, t2.value_3
  ORDER BY 1 DESC, 2 DESC, 3 DESC, 4 DESC 
  LIMIT 5;
 ERROR:  complex joins are only supported when all distributed tables are joined on their distribution columns with equal operator
+-- outer joins as subqueries should work
+-- https://github.com/citusdata/citus/issues/2739
+SELECT user_id, value_1, event_type
+FROM (
+	SELECT a.user_id, a.value_1, b.event_type
+	FROM users_table a
+	LEFT JOIN events_table b ON a.user_id = b.user_id
+) lo
+ORDER BY 1, 2, 3
+LIMIT 5;
+ user_id | value_1 | event_type 
+---------+---------+------------
+       1 |       1 |          0
+       1 |       1 |          0
+       1 |       1 |          1
+       1 |       1 |          1
+       1 |       1 |          2
+(5 rows)
+
 -- inner joins on reference tables with functions works
 SELECT DISTINCT ON (t1.user_id) t1.user_id, t2.value_1, t2.value_2, t2.value_3
 FROM events_table t1
diff --git a/src/test/regress/expected/non_colocated_subquery_joins.out b/src/test/regress/expected/non_colocated_subquery_joins.out
index dbac68871..7e2e0dd6e 100644
--- a/src/test/regress/expected/non_colocated_subquery_joins.out
+++ b/src/test/regress/expected/non_colocated_subquery_joins.out
@@ -957,7 +957,7 @@ $$);
 DEBUG:  skipping recursive planning for the subquery since it contains references to outer queries
 DEBUG:  skipping recursive planning for the subquery since it contains references to outer queries
 DEBUG:  skipping recursive planning for the subquery since it contains references to outer queries
-ERROR:  cannot perform distributed planning on this query
+ERROR:  complex joins are only supported when all distributed tables are joined on their distribution columns with equal operator
 -- similar to the above, make sure that we skip recursive plannig when
 -- the subquery contains only intermediate results
 SELECT *
diff --git a/src/test/regress/sql/multi_subquery.sql b/src/test/regress/sql/multi_subquery.sql
index 2665a1141..d65453486 100644
--- a/src/test/regress/sql/multi_subquery.sql
+++ b/src/test/regress/sql/multi_subquery.sql
@@ -314,6 +314,17 @@ SELECT DISTINCT ON (t1.user_id) t1.user_id, t2.value_1, t2.value_2, t2.value_3
  ORDER BY 1 DESC, 2 DESC, 3 DESC, 4 DESC 
  LIMIT 5;
 
+-- outer joins as subqueries should work
+-- https://github.com/citusdata/citus/issues/2739
+SELECT user_id, value_1, event_type
+FROM (
+	SELECT a.user_id, a.value_1, b.event_type
+	FROM users_table a
+	LEFT JOIN events_table b ON a.user_id = b.user_id
+) lo
+ORDER BY 1, 2, 3
+LIMIT 5;
+
 -- inner joins on reference tables with functions works
 SELECT DISTINCT ON (t1.user_id) t1.user_id, t2.value_1, t2.value_2, t2.value_3
 FROM events_table t1