Merge pull request #3878 from citusdata/explain_analyze_initial_cleanup

Explain Analyze tests & cleanup

src/backend/distributed/planner/multi_explain.c

@@ -57,13 +57,6 @@
 #include "utils/snapmgr.h"
 
 
-/* OR-able flags for ExplainXMLTag() (explain.c) */
-#define X_OPENING 0
-#define X_CLOSING 1
-#define X_CLOSE_IMMEDIATE 2
-#define X_NOWHITESPACE 4
-
-
 /* Config variables that enable printing distributed query plans */
 bool ExplainDistributedQueries = true;
 bool ExplainAllTasks = false;
@@ -148,21 +141,23 @@ CoordinatorInsertSelectExplainScan(CustomScanState *node, List *ancestors,
 	char *queryString = NULL;
 	int cursorOptions = CURSOR_OPT_PARALLEL_OK;
 
-	if (es->analyze)
-	{
-		/* avoiding double execution here is tricky, error out for now */
-		ereport(ERROR, (errmsg("EXPLAIN ANALYZE is currently not supported for INSERT "
-							   "... SELECT commands via the coordinator")));
-	}
-
 	/*
 	 * Make a copy of the query, since pg_plan_query may scribble on it and later
 	 * stages of EXPLAIN require it.
 	 */
 	Query *queryCopy = copyObject(query);
 	PlannedStmt *selectPlan = pg_plan_query(queryCopy, cursorOptions, params);
-	if (IsRedistributablePlan(selectPlan->planTree) &&
+	bool repartition = IsRedistributablePlan(selectPlan->planTree) &&
-		IsSupportedRedistributionTarget(targetRelationId))
+					   IsSupportedRedistributionTarget(targetRelationId);
+
+	if (es->analyze)
+	{
+		ereport(ERROR, (errmsg("EXPLAIN ANALYZE is currently not supported for INSERT "
+							   "... SELECT commands %s",
+							   repartition ? "with repartitioning" : "via coordinator")));
+	}
+
+	if (repartition)
 	{
 		ExplainPropertyText("INSERT/SELECT method", "repartition", es);
 	}

src/test/regress/expected/insert_select_repartition.out

@@ -573,6 +573,11 @@ EXPLAIN INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
                      ->  Seq Scan on source_table_4213606 source_table  (cost=0.00..32.60 rows=2260 width=8)
 (10 rows)
 
+--
+-- EXPLAIN ANALYZE is currently not supported
+--
+EXPLAIN ANALYZE INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
+ERROR:  EXPLAIN ANALYZE is currently not supported for INSERT ... SELECT commands  with repartitioning
 --
 -- Duplicate names in target list
 --

src/test/regress/expected/multi_explain.out

@@ -289,6 +289,27 @@ Sort (actual rows=50 loops=1)
                     ->  HashAggregate (actual rows=50 loops=1)
                           Group Key: l_quantity
                           ->  Seq Scan on lineitem_290000 lineitem (actual rows=6000 loops=1)
+-- EXPLAIN ANALYZE doesn't show worker tasks for repartition joins yet
+SET citus.shard_count TO 3;
+CREATE TABLE t1(a int, b int);
+CREATE TABLE t2(a int, b int);
+SELECT create_distributed_table('t1', 'a'), create_distributed_table('t2', 'a');
+|
+BEGIN;
+SET LOCAL citus.enable_repartition_joins TO true;
+EXPLAIN (COSTS off, ANALYZE on, TIMING off, SUMMARY off) SELECT count(*) FROM t1, t2 WHERE t1.a=t2.b;
+Aggregate (actual rows=1 loops=1)
+  ->  Custom Scan (Citus Adaptive) (actual rows=4 loops=1)
+        Task Count: 4
+        Tasks Shown: None, not supported for re-partition queries
+        ->  MapMergeJob
+              Map Task Count: 3
+              Merge Task Count: 4
+        ->  MapMergeJob
+              Map Task Count: 3
+              Merge Task Count: 4
+END;
+DROP TABLE t1, t2;
 -- Test verbose
 EXPLAIN (COSTS FALSE, VERBOSE TRUE)
 	SELECT sum(l_quantity) / avg(l_quantity) FROM lineitem;

src/test/regress/expected/multi_insert_select.out

@@ -910,6 +910,14 @@ $Q$);
          Task Count: 4
 (4 rows)
 
+-- EXPLAIN ANALYZE is not supported for INSERT ... SELECT via coordinator
+EXPLAIN (costs off, analyze on)
+ INSERT INTO agg_events (user_id)
+ SELECT
+   raw_events_first.user_id
+ FROM
+   raw_events_first INNER JOIN raw_events_second ON raw_events_first.value_1 = raw_events_second.value_1;
+ERROR:  EXPLAIN ANALYZE is currently not supported for INSERT ... SELECT commands via coordinator
 -- even if there is a filter on the partition key, since the join is not on the partition key we reject
 -- this query
 INSERT INTO agg_events (user_id)

src/test/regress/sql/insert_select_repartition.sql

@@ -271,6 +271,12 @@ SELECT * FROM target_table ORDER BY a;
 
 EXPLAIN INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
 
+--
+-- EXPLAIN ANALYZE is currently not supported
+--
+
+EXPLAIN ANALYZE INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
+
 --
 -- Duplicate names in target list
 --

src/test/regress/sql/multi_explain.sql

@@ -85,6 +85,17 @@ EXPLAIN (COSTS FALSE, ANALYZE TRUE, TIMING FALSE, SUMMARY FALSE)
 	SELECT l_quantity, count(*) count_quantity FROM lineitem
 	GROUP BY l_quantity ORDER BY count_quantity, l_quantity;
 
+-- EXPLAIN ANALYZE doesn't show worker tasks for repartition joins yet
+SET citus.shard_count TO 3;
+CREATE TABLE t1(a int, b int);
+CREATE TABLE t2(a int, b int);
+SELECT create_distributed_table('t1', 'a'), create_distributed_table('t2', 'a');
+BEGIN;
+SET LOCAL citus.enable_repartition_joins TO true;
+EXPLAIN (COSTS off, ANALYZE on, TIMING off, SUMMARY off) SELECT count(*) FROM t1, t2 WHERE t1.a=t2.b;
+END;
+DROP TABLE t1, t2;
+
 -- Test verbose
 EXPLAIN (COSTS FALSE, VERBOSE TRUE)
 	SELECT sum(l_quantity) / avg(l_quantity) FROM lineitem;

src/test/regress/sql/multi_insert_select.sql

@@ -677,6 +677,14 @@ SET client_min_messages TO WARNING;
    raw_events_first INNER JOIN raw_events_second ON raw_events_first.value_1 = raw_events_second.value_1;
 $Q$);
 
+-- EXPLAIN ANALYZE is not supported for INSERT ... SELECT via coordinator
+EXPLAIN (costs off, analyze on)
+ INSERT INTO agg_events (user_id)
+ SELECT
+   raw_events_first.user_id
+ FROM
+   raw_events_first INNER JOIN raw_events_second ON raw_events_first.value_1 = raw_events_second.value_1;
+
 -- even if there is a filter on the partition key, since the join is not on the partition key we reject
 -- this query
 INSERT INTO agg_events (user_id)