Strip trailing whitespace and add final newline (#3186)

This brings files in line with our editorconfig file

.circleci/config.yml

@@ -22,6 +22,12 @@ jobs:
       - run:
           name: 'Check Style'
           command: citus_indent --check
+      - run:
+          name: 'Fix whitespace'
+          command: ci/editorconfig.sh
+      - run:
+          name: 'Check if whitespace fixing changed anything, install editorconfig if it did'
+          command: git diff --exit-code
       - run:
           name: 'Remove useless declarations'
           command: ci/remove_useless_declarations.sh

.editorconfig

@@ -12,12 +12,18 @@ insert_final_newline = true
 charset = utf-8
 trim_trailing_whitespace = true
 
-# Don't change test output files
-[*.out]
+# Don't change test output files, pngs or test data files
+[*.{out,png,data}]
 insert_final_newline = unset
 trim_trailing_whitespace = unset
 
-[*.sql]
+# Don't change test/regress/output directory, this needs to be a separate rule
+# for some reason
+[/src/test/regress/output/**]
+insert_final_newline = unset
+trim_trailing_whitespace = unset
+
+[*.sql,*.sh]
 indent_style = space
 indent_size = 4
 tab_width = 4

LICENSE

@@ -658,4 +658,4 @@ specific requirements.
   You should also get your employer (if you work as a programmer) or school,
 if any, to sign a "copyright disclaimer" for the program, if necessary.
 For more information on this, and how to apply and follow the GNU AGPL, see
-<http://www.gnu.org/licenses/>.
+<http://www.gnu.org/licenses/>.

ci/editorconfig.sh

@@ -0,0 +1,16 @@
+#!/bin/sh
+set -eu
+for f in $(git ls-tree -r HEAD --name-only); do
+    if [ "$f" = "${f%.out}" ]  &&
+        [ "$f" = "${f%.data}" ] &&
+        [ "$f" = "${f%.png}" ] &&
+        [ "$(dirname "$f")" != "src/test/regress/output" ]
+    then
+        # Trim trailing whitespace
+        sed -e 's/[[:space:]]*$//' -i "./$f"
+        # Add final newline if not there
+        if [ -n "$(tail -c1 "$f")" ]; then
+            echo >> "$f"
+        fi
+    fi
+done

src/backend/distributed/planner/planner_readme.md

@@ -6,7 +6,7 @@ If the input query is trivial (e.g., no joins, no subqueries/ctes, single table
 
 Distributed planning (`CreateDistributedPlan`) tries several different methods to plan the query:
 
- 
+
  1. Fast-path router planner, proceed if the query prunes down to a single shard of a single table
  2. Router planner, proceed if the query prunes down to a single set of co-located shards
  3. Modification planning, proceed if the query is a DML command and all joins are co-located
@@ -15,7 +15,7 @@ Distributed planning (`CreateDistributedPlan`) tries several different methods t
 
 ## Fast-path router planner
 
-By examining the query tree, if we can decide that the query hits only a single shard of a single table, we can skip calling `standard_planner()`. Later on the execution, we simply fetch the filter on the distribution key and do the pruning. 
+By examining the query tree, if we can decide that the query hits only a single shard of a single table, we can skip calling `standard_planner()`. Later on the execution, we simply fetch the filter on the distribution key and do the pruning.
 
 As the name reveals, this can be considered as a sub-item of Router planner described below. The only difference is that fast-path planner doesn't rely on `standard_planner()` for collecting restriction information.
 

src/backend/distributed/sql/citus--7.0-1--7.0-2.sql

@@ -1,6 +1,6 @@
---  citus--7.0-1--7.0-2.sql 
+--  citus--7.0-1--7.0-2.sql
 
---  redefine shard_name as STRICT 
+--  redefine shard_name as STRICT
 CREATE OR REPLACE FUNCTION pg_catalog.shard_name(object_name regclass, shard_id bigint)
     RETURNS text
     LANGUAGE C STABLE STRICT

src/backend/distributed/sql/citus--7.0-10--7.0-11.sql

@@ -1,4 +1,4 @@
---  citus-7.0-10--7.0-11 
+--  citus-7.0-10--7.0-11
 
 SET search_path = 'pg_catalog';
 
@@ -8,9 +8,9 @@ DECLARE
 	colocated_tables regclass[];
 BEGIN
 	SELECT get_colocated_table_array(relation) INTO colocated_tables;
-	
-	PERFORM 
-		master_update_shard_statistics(shardid) 
+
+	PERFORM
+		master_update_shard_statistics(shardid)
 	FROM
 		pg_dist_shard
 	WHERE
@@ -19,7 +19,7 @@ END;
 $$ LANGUAGE 'plpgsql';
 COMMENT ON FUNCTION master_update_table_statistics(regclass)
 	IS 'updates shard statistics of the given table and its colocated tables';
-	
+
 CREATE OR REPLACE FUNCTION get_colocated_shard_array(bigint)
 	RETURNS BIGINT[]
 	LANGUAGE C STRICT

src/backend/distributed/sql/citus--7.0-11--7.0-12.sql

@@ -1,4 +1,4 @@
---  citus--7.0-11--7.0-12.sql 
+--  citus--7.0-11--7.0-12.sql
 
 CREATE OR REPLACE FUNCTION pg_catalog.citus_create_restore_point(text)
 RETURNS pg_lsn

src/backend/distributed/sql/citus--7.0-12--7.0-13.sql

@@ -1,4 +1,4 @@
---  citus--7.0-12--7.0-13.sql 
+--  citus--7.0-12--7.0-13.sql
 
 SET search_path = 'pg_catalog';
 
@@ -18,7 +18,7 @@ BEGIN
     SELECT array_agg(object_identity) INTO sequence_names
     FROM pg_event_trigger_dropped_objects()
     WHERE object_type = 'sequence';
-   
+
     FOR v_obj IN SELECT * FROM pg_event_trigger_dropped_objects() JOIN
                                pg_dist_partition ON (logicalrelid = objid)
                  WHERE object_type IN ('table', 'foreign table')
@@ -28,20 +28,20 @@ BEGIN
 
         -- ensure all shards are dropped
         PERFORM master_drop_all_shards(v_obj.objid, v_obj.schema_name, v_obj.object_name);
-        
+
         PERFORM master_drop_distributed_table_metadata(v_obj.objid, v_obj.schema_name, v_obj.object_name);
     END LOOP;
 
     IF cardinality(sequence_names) = 0 THEN
         RETURN;
     END IF;
-    
+
     PERFORM master_drop_sequences(sequence_names);
 END;
 $cdbdt$;
 
 COMMENT ON FUNCTION citus_drop_trigger()
     IS 'perform checks and actions at the end of DROP actions';
-    
+
 RESET search_path;
 

src/backend/distributed/sql/citus--7.0-13--7.0-14.sql

@@ -1,4 +1,4 @@
---  citus--7.0-13--7.0-14.sql 
+--  citus--7.0-13--7.0-14.sql
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-14--7.0-15.sql

@@ -1,4 +1,4 @@
---  citus--7.0-14--7.0-15 
+--  citus--7.0-14--7.0-15
 
 DROP FUNCTION pg_catalog.dump_local_wait_edges(int4);
 

src/backend/distributed/sql/citus--7.0-15--7.1-1.sql

@@ -1,4 +1,4 @@
---  citus--7.0-15--7.1-1 
+--  citus--7.0-15--7.1-1
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-2--7.0-3.sql

@@ -1,4 +1,4 @@
---  citus--7.0-2--7.0-3.sql 
+--  citus--7.0-2--7.0-3.sql
 
 ALTER SEQUENCE pg_catalog.pg_dist_shard_placement_placementid_seq
   RENAME TO pg_dist_placement_placementid_seq;

src/backend/distributed/sql/citus--7.0-3--7.0-4.sql

@@ -1,4 +1,4 @@
---  citus--7.0-3--7.0-4.sql 
+--  citus--7.0-3--7.0-4.sql
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-4--7.0-5.sql

@@ -1,4 +1,4 @@
---  citus--7.0-4--7.0-5.sql 
+--  citus--7.0-4--7.0-5.sql
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-5--7.0-6.sql

@@ -1,4 +1,4 @@
---  citus--7.0-5--7.0-6 
+--  citus--7.0-5--7.0-6
 
 CREATE FUNCTION pg_catalog.dump_local_wait_edges(
                     IN source_node_id int4,

src/backend/distributed/sql/citus--7.0-6--7.0-7.sql

@@ -1,4 +1,4 @@
---  citus--7.0-6--7.0-7 
+--  citus--7.0-6--7.0-7
 
 CREATE FUNCTION citus.replace_isolation_tester_func()
 RETURNS void AS $$
@@ -66,7 +66,7 @@ RETURNS int4[] AS $$
     -- pg says we're not blocked locally; check whether we're blocked globally.
     SELECT transaction_number INTO mLocalTransactionNum
       FROM get_all_active_transactions() WHERE process_id = pBlockedPid;
-    
+
     SELECT array_agg(process_id) INTO mRemoteBlockingPids FROM (
       WITH activeTransactions AS (
         SELECT process_id, transaction_number FROM get_all_active_transactions()

src/backend/distributed/sql/citus--7.0-7--7.0-8.sql

@@ -1,4 +1,4 @@
---  citus--7.0-7--7.0-8.sql 
+--  citus--7.0-7--7.0-8.sql
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-8--7.0-9.sql

@@ -1,4 +1,4 @@
---  citus-7.0-8--7.0-9 
+--  citus-7.0-8--7.0-9
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.0-9--7.0-10.sql

@@ -1,4 +1,4 @@
---  citus-7.0-9--7.0-10 
+--  citus-7.0-9--7.0-10
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--7.1-1--7.1-2.sql

@@ -1,4 +1,4 @@
---  citus--7.1-1--7.1-2 
+--  citus--7.1-1--7.1-2
 
 CREATE OR REPLACE FUNCTION pg_catalog.citus_version()
     RETURNS text

src/backend/distributed/sql/citus--7.1-2--7.1-3.sql

@@ -1,4 +1,4 @@
---  citus--7.1-2--7.1-3 
+--  citus--7.1-2--7.1-3
 
 CREATE TABLE citus.pg_dist_node_metadata(
     metadata jsonb NOT NULL

src/backend/distributed/sql/citus--7.1-3--7.1-4.sql

@@ -1,4 +1,4 @@
---  citus--7.1-3--7.1-4 
+--  citus--7.1-3--7.1-4
 
 CREATE TYPE citus.shard_transfer_mode AS ENUM (
    'auto',

src/backend/distributed/sql/citus--7.1-4--7.2-1.sql

@@ -1,4 +1,4 @@
---  citus--7.1-4--7.2-1 
+--  citus--7.1-4--7.2-1
 
---  bump version to 7.2-1 
+--  bump version to 7.2-1
 

src/backend/distributed/sql/citus--7.2-1--7.2-2.sql

@@ -1,4 +1,4 @@
---  citus--7.2-1--7.2-2 
+--  citus--7.2-1--7.2-2
 
 CREATE TYPE citus.copy_format AS ENUM ('csv', 'binary', 'text');
 

src/backend/distributed/sql/citus--7.2-2--7.2-3.sql

@@ -1,4 +1,4 @@
---  citus--7.2-2--7.2-3 
+--  citus--7.2-2--7.2-3
 
 DROP FUNCTION pg_catalog.read_intermediate_result(text,citus.copy_format);
 DROP TYPE citus.copy_format;

src/backend/distributed/sql/citus--7.2-3--7.3-1.sql

@@ -1,4 +1,4 @@
---  citus--7.2-3--7.3-1 
+--  citus--7.2-3--7.3-1
 
---  bump version to 7.3-1 
+--  bump version to 7.3-1
 

src/backend/distributed/sql/citus--7.3-1--7.3-2.sql

@@ -1,4 +1,4 @@
---  citus--7.3-1--7.3-2 
+--  citus--7.3-1--7.3-2
 
 CREATE FUNCTION pg_catalog.citus_text_send_as_jsonb(text)
 RETURNS bytea

src/backend/distributed/sql/citus--7.3-2--7.3-3.sql

@@ -1,4 +1,4 @@
---  citus--7.3-2--7.3-3 
+--  citus--7.3-2--7.3-3
 
 -- Citus json aggregate helpers
 
@@ -6,7 +6,7 @@ CREATE FUNCTION pg_catalog.citus_jsonb_concatenate(state jsonb, val jsonb)
 	RETURNS jsonb
 	LANGUAGE SQL
 AS $function$
-	SELECT CASE 
+	SELECT CASE
 		WHEN val IS NULL THEN state
 		WHEN jsonb_typeof(state) = 'null' THEN val
 		ELSE state || val
@@ -24,7 +24,7 @@ CREATE FUNCTION pg_catalog.citus_json_concatenate(state json, val json)
 	RETURNS json
 	LANGUAGE SQL
 AS $function$
-	SELECT CASE 
+	SELECT CASE
 		WHEN val IS NULL THEN state
 		WHEN json_typeof(state) = 'null' THEN val
 		WHEN json_typeof(state) = 'object' THEN
@@ -33,7 +33,7 @@ AS $function$
 		 		UNION ALL
 		 		SELECT * FROM json_each(val)
 	 		) t)
-		ELSE 
+		ELSE
 	 		(SELECT json_agg(a) FROM (
 	 			SELECT json_array_elements(state) AS a
 		 		UNION ALL
@@ -60,7 +60,7 @@ CREATE AGGREGATE pg_catalog.jsonb_cat_agg(jsonb) (
 );
 COMMENT ON AGGREGATE pg_catalog.jsonb_cat_agg(jsonb)
     IS 'concatenate input jsonbs into a single jsonb';
-    
+
 CREATE AGGREGATE pg_catalog.json_cat_agg(json) (
     SFUNC = citus_json_concatenate,
     FINALFUNC = citus_json_concatenate_final,

src/backend/distributed/sql/citus--7.3-3--7.4-1.sql

@@ -1,4 +1,4 @@
---  citus--7.3-3--7.4-1 
+--  citus--7.3-3--7.4-1
 
 DROP FUNCTION IF EXISTS master_expire_table_cache(regclass);
 DROP FUNCTION IF EXISTS pg_catalog.worker_fetch_regular_table(text, bigint, text[], integer[]);

src/backend/distributed/sql/citus--7.4-1--7.4-2.sql

@@ -1,4 +1,4 @@
---  citus--7.4-1--7.4-2 
+--  citus--7.4-1--7.4-2
 
 CREATE OR REPLACE FUNCTION pg_catalog.citus_drop_trigger()
     RETURNS event_trigger

src/backend/distributed/sql/citus--7.4-2--7.4-3.sql

@@ -1,4 +1,4 @@
---  citus--7.4-2--7.4-3 
+--  citus--7.4-2--7.4-3
 SET search_path = 'pg_catalog';
 
 -- note that we're not dropping the older version of the function

src/backend/distributed/sql/citus--7.4-3--7.5-1.sql

@@ -1,4 +1,4 @@
---  citus--7.4-3--7.5-1 
+--  citus--7.4-3--7.5-1
 
---  bump version to 7.5-1 
+--  bump version to 7.5-1
 

src/backend/distributed/sql/citus--7.5-1--7.5-2.sql

@@ -1,4 +1,4 @@
---  citus--7.5-1--7.5-2 
+--  citus--7.5-1--7.5-2
 SET search_path = 'pg_catalog';
 
 -- note that we're not dropping the older version of the function

src/backend/distributed/sql/citus--7.5-2--7.5-3.sql

@@ -1,4 +1,4 @@
---  citus--7.5-2--7.5-3 
+--  citus--7.5-2--7.5-3
 SET search_path = 'pg_catalog';
 
 CREATE FUNCTION master_dist_authinfo_cache_invalidate()

src/backend/distributed/sql/citus--7.5-3--7.5-4.sql

@@ -1,4 +1,4 @@
---  citus--7.5-3--7.5-4 
+--  citus--7.5-3--7.5-4
 
 CREATE FUNCTION pg_catalog.citus_query_stats(OUT queryid bigint,
 											 OUT userid oid,

src/backend/distributed/sql/citus--7.5-4--7.5-5.sql

@@ -1,4 +1,4 @@
---  citus--7.5-4--7.5-5 
+--  citus--7.5-4--7.5-5
 CREATE FUNCTION pg_catalog.citus_executor_name(executor_type int)
 RETURNS TEXT
 LANGUAGE plpgsql

src/backend/distributed/sql/citus--7.5-5--7.5-6.sql

@@ -1,4 +1,4 @@
---  citus--7.5-5--7.5-6 
+--  citus--7.5-5--7.5-6
 
 -- Don't want this to be available to non-superusers.
 REVOKE ALL ON FUNCTION pg_catalog.citus_stat_statements_reset() FROM PUBLIC;

src/backend/distributed/sql/citus--7.5-6--7.5-7.sql

@@ -1,4 +1,4 @@
---  citus--7.5-6--7.5-7 
+--  citus--7.5-6--7.5-7
 SET search_path = 'pg_catalog';
 
 CREATE FUNCTION pg_catalog.poolinfo_valid(text)

src/backend/distributed/sql/citus--7.5-7--8.0-1.sql

@@ -1,4 +1,4 @@
---  citus--7.5-7--8.0-1 
+--  citus--7.5-7--8.0-1
 
---  bump version to 8.0-1 
+--  bump version to 8.0-1
 

src/backend/distributed/sql/citus--8.0-1--8.0-2.sql

@@ -1,4 +1,4 @@
---  citus--7.5-7--8.0-1 
+--  citus--7.5-7--8.0-1
 SET search_path = 'pg_catalog';
 
 CREATE OR REPLACE FUNCTION pg_catalog.relation_is_a_known_shard(regclass)
@@ -20,7 +20,7 @@ COMMENT ON FUNCTION citus_table_is_visible(oid)
 -- this is the exact same query with what \d
 -- command produces, except pg_table_is_visible
 -- is replaced with pg_catalog.relation_is_a_known_shard(c.oid)
-CREATE VIEW citus.citus_shards_on_worker AS 
+CREATE VIEW citus.citus_shards_on_worker AS
 	SELECT n.nspname as "Schema",
 	  c.relname as "Name",
 	  CASE c.relkind WHEN 'r' THEN 'table' WHEN 'v' THEN 'view' WHEN 'm' THEN 'materialized view' WHEN 'i' THEN 'index' WHEN 'S' THEN 'sequence' WHEN 's' THEN 'special' WHEN 'f' THEN 'foreign table' WHEN 'p' THEN 'table' END as "Type",
@@ -39,7 +39,7 @@ GRANT SELECT ON pg_catalog.citus_shards_on_worker TO public;
 -- this is the exact same query with what \di
 -- command produces, except pg_table_is_visible
 -- is replaced with pg_catalog.relation_is_a_known_shard(c.oid)
-CREATE VIEW citus.citus_shard_indexes_on_worker AS 
+CREATE VIEW citus.citus_shard_indexes_on_worker AS
 SELECT n.nspname as "Schema",
   c.relname as "Name",
   CASE c.relkind WHEN 'r' THEN 'table' WHEN 'v' THEN 'view' WHEN 'm' THEN 'materialized view' WHEN 'i' THEN 'index' WHEN 'S' THEN 'sequence' WHEN 's' THEN 'special' WHEN 'f' THEN 'foreign table' WHEN 'p' THEN 'table' END as "Type",

src/backend/distributed/sql/citus--8.0-10--8.0-11.sql

@@ -1,4 +1,4 @@
---  citus--8.0-10--8.0-11 
+--  citus--8.0-10--8.0-11
 SET search_path = 'pg_catalog';
 
 -- Deprecated functions

src/backend/distributed/sql/citus--8.0-11--8.0-12.sql

@@ -1,4 +1,4 @@
---  citus--8.0-11--8.0-12 
+--  citus--8.0-11--8.0-12
 SET search_path = 'pg_catalog';
 
 CREATE OR REPLACE FUNCTION pg_catalog.citus_stat_statements(OUT queryid bigint,

src/backend/distributed/sql/citus--8.0-12--8.0-13.sql

@@ -1,4 +1,4 @@
---  citus--8.0-12--8.0-13 
+--  citus--8.0-12--8.0-13
 CREATE FUNCTION citus_check_defaults_for_sslmode()
     RETURNS void
     LANGUAGE C STRICT

src/backend/distributed/sql/citus--8.0-13--8.1-1.sql

@@ -1,4 +1,4 @@
---  citus--8.0-13--8.1-1.sql 
+--  citus--8.0-13--8.1-1.sql
 
---  bump version to 8.1-1 
+--  bump version to 8.1-1
 

src/backend/distributed/sql/citus--8.0-2--8.0-3.sql

@@ -1,4 +1,4 @@
---  citus--8.0-2--8.0-3 
+--  citus--8.0-2--8.0-3
 SET search_path = 'pg_catalog';
 
 CREATE FUNCTION master_remove_partition_metadata(logicalrelid regclass,

src/backend/distributed/sql/citus--8.0-3--8.0-4.sql

@@ -1,4 +1,4 @@
---  citus--8.0-3--8.0-4 
+--  citus--8.0-3--8.0-4
 SET search_path = 'pg_catalog';
 
 CREATE OR REPLACE FUNCTION lock_relation_if_exists(table_name text, lock_mode text)

src/backend/distributed/sql/citus--8.0-4--8.0-5.sql

@@ -1,37 +1,37 @@
---  citus--8.0-4--8.0-5.sql 
+--  citus--8.0-4--8.0-5.sql
 SET search_path = 'pg_catalog';
 
 
 DROP FUNCTION IF EXISTS get_all_active_transactions();
 
 
-CREATE OR REPLACE FUNCTION get_all_active_transactions(OUT datid oid, OUT process_id int, OUT initiator_node_identifier int4, OUT worker_query BOOL, 
-                                                       OUT transaction_number int8, OUT transaction_stamp timestamptz) 
-RETURNS SETOF RECORD 
-LANGUAGE C STRICT AS 'MODULE_PATHNAME', 
+CREATE OR REPLACE FUNCTION get_all_active_transactions(OUT datid oid, OUT process_id int, OUT initiator_node_identifier int4, OUT worker_query BOOL,
+                                                       OUT transaction_number int8, OUT transaction_stamp timestamptz)
+RETURNS SETOF RECORD
+LANGUAGE C STRICT AS 'MODULE_PATHNAME',
 $$get_all_active_transactions$$;
 
-COMMENT ON FUNCTION get_all_active_transactions(OUT datid oid, OUT datname text, OUT process_id int, OUT initiator_node_identifier int4, OUT worker_query BOOL, 
-                                                OUT transaction_number int8, OUT transaction_stamp timestamptz) 
+COMMENT ON FUNCTION get_all_active_transactions(OUT datid oid, OUT datname text, OUT process_id int, OUT initiator_node_identifier int4, OUT worker_query BOOL,
+                                                OUT transaction_number int8, OUT transaction_stamp timestamptz)
 IS 'returns distributed transaction ids of active distributed transactions';
 
 
-CREATE OR REPLACE FUNCTION citus_dist_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int, 
-                                                    OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name, 
-                                                    OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET, 
-                                                    OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz, 
-                                                    OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text, 
-                                                    OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text) 
-RETURNS SETOF RECORD 
-LANGUAGE C STRICT AS 'MODULE_PATHNAME', 
+CREATE OR REPLACE FUNCTION citus_dist_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int,
+                                                    OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name,
+                                                    OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET,
+                                                    OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz,
+                                                    OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text,
+                                                    OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text)
+RETURNS SETOF RECORD
+LANGUAGE C STRICT AS 'MODULE_PATHNAME',
 $$citus_dist_stat_activity$$;
 
-COMMENT ON FUNCTION citus_dist_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int, 
-                                             OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name, 
-                                             OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET, 
-                                             OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz, 
-                                             OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text, 
-                                             OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text) 
+COMMENT ON FUNCTION citus_dist_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int,
+                                             OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name,
+                                             OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET,
+                                             OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz,
+                                             OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text,
+                                             OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text)
 IS 'returns distributed transaction activity on distributed tables';
 
 CREATE VIEW citus.citus_dist_stat_activity AS
@@ -41,22 +41,22 @@ GRANT SELECT ON pg_catalog.citus_dist_stat_activity TO PUBLIC;
 
 
 
-CREATE OR REPLACE FUNCTION citus_worker_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int, 
-                                                      OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name, 
-                                                      OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET, 
-                                                      OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz, 
-                                                      OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text, 
-                                                      OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text) 
-RETURNS SETOF RECORD 
-LANGUAGE C STRICT AS 'MODULE_PATHNAME', 
+CREATE OR REPLACE FUNCTION citus_worker_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int,
+                                                      OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name,
+                                                      OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET,
+                                                      OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz,
+                                                      OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text,
+                                                      OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text)
+RETURNS SETOF RECORD
+LANGUAGE C STRICT AS 'MODULE_PATHNAME',
 $$citus_worker_stat_activity$$;
 
-COMMENT ON FUNCTION citus_worker_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int, 
-                                               OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name, 
-                                               OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET, 
-                                               OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz, 
-                                               OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text, 
-                                               OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text) 
+COMMENT ON FUNCTION citus_worker_stat_activity(OUT query_hostname text, OUT query_hostport int, OUT master_query_host_name text, OUT master_query_host_port int,
+                                               OUT transaction_number int8, OUT transaction_stamp timestamptz, OUT datid oid, OUT datname name,
+                                               OUT pid int, OUT usesysid oid, OUT usename name, OUT application_name text, OUT client_addr INET,
+                                               OUT client_hostname TEXT, OUT client_port int, OUT backend_start timestamptz, OUT xact_start timestamptz,
+                                               OUT query_start timestamptz, OUT state_change timestamptz, OUT wait_event_type text, OUT wait_event text,
+                                               OUT state text, OUT backend_xid xid, OUT backend_xmin xid, OUT query text, OUT backend_type text)
 IS 'returns distributed transaction activity on shards of distributed tables';
 
 CREATE VIEW citus.citus_worker_stat_activity AS

src/backend/distributed/sql/citus--8.0-5--8.0-6.sql

@@ -1,4 +1,4 @@
---  citus--8.0-5--8.0-6 
+--  citus--8.0-5--8.0-6
 SET search_path = 'pg_catalog';
 
 CREATE FUNCTION get_global_active_transactions(OUT datid oid, OUT process_id int, OUT initiator_node_identifier int4, OUT worker_query BOOL, OUT transaction_number int8, OUT transaction_stamp timestamptz)
@@ -42,10 +42,10 @@ RETURNS int4[] AS $$
                  WHERE waiting_transaction_num = mLocalTransactionNum) THEN
       SELECT array_agg(pBlockedPid) INTO mRemoteBlockingPids;
     END IF;
-    
+
     RETURN mRemoteBlockingPids;
   END;
-$$ LANGUAGE plpgsql;    
+$$ LANGUAGE plpgsql;
 
 #include "udfs/citus_isolation_test_session_is_blocked/8.0-6.sql"
 

src/backend/distributed/sql/citus--8.0-6--8.0-7.sql

@@ -1,9 +1,9 @@
---  citus--8.0-6--8.0-7 
+--  citus--8.0-6--8.0-7
 SET search_path = 'pg_catalog';
 
 CREATE VIEW citus.citus_lock_waits AS
 
-WITH 
+WITH
 citus_dist_stat_activity AS
 (
   SELECT * FROM citus_dist_stat_activity
@@ -19,7 +19,7 @@ citus_dist_stat_activity_with_node_id AS
   FROM
   citus_dist_stat_activity LEFT JOIN pg_dist_node
   ON
-  citus_dist_stat_activity.master_query_host_name = pg_dist_node.nodename AND 
+  citus_dist_stat_activity.master_query_host_name = pg_dist_node.nodename AND
   citus_dist_stat_activity.master_query_host_port = pg_dist_node.nodeport
 )
 SELECT

src/backend/distributed/sql/citus--8.0-7--8.0-8.sql

@@ -1,14 +1,14 @@
---  citus--8.0-7--8.0-8 
+--  citus--8.0-7--8.0-8
 SET search_path = 'pg_catalog';
 
 DROP FUNCTION IF EXISTS pg_catalog.worker_drop_distributed_table(logicalrelid Oid);
 
-	
+
 CREATE FUNCTION worker_drop_distributed_table(table_name text)
     RETURNS VOID
     LANGUAGE C STRICT
     AS 'MODULE_PATHNAME', $$worker_drop_distributed_table$$;
-    
+
 COMMENT ON FUNCTION worker_drop_distributed_table(table_name text)
     IS 'drop the distributed table and its reference from metadata tables';
 

src/backend/distributed/sql/citus--8.0-9--8.0-10.sql

@@ -1,4 +1,4 @@
---  citus--8.0-9--8.0-10 
+--  citus--8.0-9--8.0-10
 SET search_path = 'pg_catalog';
 
 CREATE FUNCTION worker_execute_sql_task(jobid bigint, taskid integer, query text, binary bool)

src/backend/distributed/sql/citus--8.1-1--8.2-1.sql

@@ -1,4 +1,4 @@
---  citus--8.1-1--8.2-1.sql 
+--  citus--8.1-1--8.2-1.sql
 
---  bump version to 8.2-1 
+--  bump version to 8.2-1
 

src/backend/distributed/sql/citus--8.2-1--8.2-2.sql

@@ -1,3 +1,3 @@
---  citus--8.2-1--8.2-2.sql 
+--  citus--8.2-1--8.2-2.sql
 
 DROP FUNCTION IF EXISTS pg_catalog.create_insert_proxy_for_table(regclass,regclass);

src/backend/distributed/sql/citus--8.2-2--8.2-3.sql

@@ -1,4 +1,4 @@
---  citus--8.2-2--8.2-3 
+--  citus--8.2-2--8.2-3
 
 SET search_path = 'pg_catalog';
 

src/backend/distributed/sql/citus--8.2-3--8.2-4.sql

@@ -1,4 +1,4 @@
---  citus--8.2-3--8.2-4 
+--  citus--8.2-3--8.2-4
 
 CREATE OR REPLACE FUNCTION pg_catalog.citus_executor_name(executor_type int)
 RETURNS text

src/backend/distributed/sql/citus--8.2-4--8.3-1.sql

@@ -1,4 +1,4 @@
---  citus--8.2-4--8.3-1 
+--  citus--8.2-4--8.3-1
 
---  bump version to 8.3-1 
+--  bump version to 8.3-1
 

src/backend/distributed/sql/citus--8.3-1--9.0-1.sql

@@ -1,4 +1,4 @@
---  citus--8.3-1--9.0-1 
+--  citus--8.3-1--9.0-1
 
 SET search_path = 'pg_catalog';
 
@@ -77,7 +77,7 @@ CREATE TRIGGER dist_object_cache_invalidate
 -- by the operator.
 ALTER TABLE pg_dist_poolinfo DROP CONSTRAINT pg_dist_poolinfo_nodeid_fkey;
 
---  if the rebalancer extension is still around, drop it before creating Citus functions 
+--  if the rebalancer extension is still around, drop it before creating Citus functions
 DROP EXTENSION IF EXISTS shard_rebalancer;
 
 #include "udfs/get_rebalance_table_shards_plan/9.0-1.sql"

src/backend/distributed/sql/udfs/alter_role_if_exists/9.1-1.sql

@@ -9,4 +9,3 @@ COMMENT ON FUNCTION pg_catalog.alter_role_if_exists(
     role_name text,
     utility_query text)
     IS 'runs the utility query, if the role exists';
-    

src/backend/distributed/sql/udfs/alter_role_if_exists/latest.sql

@@ -9,4 +9,3 @@ COMMENT ON FUNCTION pg_catalog.alter_role_if_exists(
     role_name text,
     utility_query text)
     IS 'runs the utility query, if the role exists';
-    

src/backend/distributed/sql/udfs/citus_isolation_test_session_is_blocked/8.0-6.sql

@@ -15,7 +15,7 @@ RETURNS boolean AS $$
     -- number when the worker process waiting for other session.
     IF EXISTS (SELECT transaction_number FROM get_global_active_transactions()
                WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId) THEN
-      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions() 
+      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions()
       WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId;
     ELSE
       -- Check whether transactions initiated from the coordinator get locked

src/backend/distributed/sql/udfs/citus_isolation_test_session_is_blocked/9.0-1.sql

@@ -15,7 +15,7 @@ RETURNS boolean AS $$
     -- number when the worker process waiting for other session.
     IF EXISTS (SELECT transaction_number FROM get_global_active_transactions()
                WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId) THEN
-      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions() 
+      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions()
       WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId;
     ELSE
       -- Check whether transactions initiated from the coordinator get locked

src/backend/distributed/sql/udfs/citus_isolation_test_session_is_blocked/latest.sql

@@ -15,7 +15,7 @@ RETURNS boolean AS $$
     -- number when the worker process waiting for other session.
     IF EXISTS (SELECT transaction_number FROM get_global_active_transactions()
                WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId) THEN
-      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions() 
+      SELECT transaction_number INTO mBlockedTransactionNum FROM get_global_active_transactions()
       WHERE process_id = workerProcessId AND pBlockedPid = coordinatorProcessId;
     ELSE
       -- Check whether transactions initiated from the coordinator get locked

src/test/regress/Makefile

@@ -165,14 +165,14 @@ check-citus-upgrade-local:
 	$(citus_upgrade_check) \
 		--bindir=$(bindir) \
 		--pgxsdir=$(pgxsdir) \
-		--citus-old-version=$(citus-old-version)	
+		--citus-old-version=$(citus-old-version)
 
 check-citus-upgrade-mixed-local:
 	$(citus_upgrade_check) \
 		--bindir=$(bindir) \
 		--pgxsdir=$(pgxsdir) \
 		--citus-old-version=$(citus-old-version) \
-		--mixed			
+		--mixed
 
 clean distclean maintainer-clean:
 	rm -f $(output_files) $(input_files)

src/test/regress/expected/cte_nested_modification.out

@@ -30,7 +30,6 @@ WITH cte_1 AS (
         FROM tt1
         WHERE value_1 >= 2
     )
-    
     DELETE FROM tt2
     USING cte_2
     WHERE tt2.id = cte_2.cte2_id
@@ -57,7 +56,6 @@ WITH cte_1 AS (
         FROM tt1
         WHERE value_1 >= 2
     )
-    
     DELETE FROM tt2
     USING cte_2
     WHERE tt2.id = cte_2.cte2_id
@@ -82,9 +80,8 @@ WITH cte_1(id) AS (
         FROM tt1
         WHERE value_1 >= 2
     )
-    
     DELETE FROM tt2
-    USING cte_2 
+    USING cte_2
     WHERE tt2.id = cte_2.cte2_id
     RETURNING cte2_id
 )
@@ -107,7 +104,6 @@ WITH cte_1 AS (
         FROM tt1
         WHERE value_1 >= 2
     )
-    
     UPDATE tt2
     SET value_1 = 10
     FROM cte_2
@@ -130,7 +126,6 @@ WITH cte_1 AS (
     WITH cte_2 AS (
         SELECT * FROM tt3
     )
-    
     UPDATE tt2
     SET value_1 = (SELECT max((json_val->>'qty')::int) FROM cte_2)
     RETURNING id, value_1

src/test/regress/expected/dml_recursive.out

@@ -27,25 +27,25 @@ INSERT INTO distributed_table SELECT i::text, i % 10, row_to_json(row(i, i*i)) F
 INSERT INTO second_distributed_table SELECT i::text, i % 10, row_to_json(row(i, i*i)) FROM generate_series (0, 100) i;
 INSERT INTO reference_table SELECT i::text, 'user_' || i FROM generate_series (0, 100) i;
 INSERT INTO local_table SELECT i::text, 'user_' || i FROM generate_series (0, 100) i;
-CREATE VIEW tenant_ids AS 
-	SELECT 
-		tenant_id, name 
-	FROM 
+CREATE VIEW tenant_ids AS
+	SELECT
+		tenant_id, name
+	FROM
 		distributed_table, reference_table
-	WHERE 
+	WHERE
 		distributed_table.dept::text = reference_table.id
 	ORDER BY 2 DESC, 1 DESC;
 SET client_min_messages TO DEBUG1;
 -- the subquery foo is recursively planned
-UPDATE 
-	reference_table 
-SET 
-	name = 'new_' || name 
-FROM 
+UPDATE
+	reference_table
+SET
+	name = 'new_' || name
+FROM
 (
-	SELECT 
+	SELECT
 		avg(second_distributed_table.tenant_id::int) as avg_tenant_id
-	FROM 
+	FROM
 		second_distributed_table
 ) as foo
 WHERE
@@ -61,27 +61,27 @@ DEBUG:  Plan 4 query after replacing subqueries and CTEs: UPDATE recursive_dml_q
 
 -- the subquery foo is recursively planned
 -- but note that the subquery foo itself is pushdownable
-UPDATE 
-	second_distributed_table 
-SET 
+UPDATE
+	second_distributed_table
+SET
 	dept = foo.max_dept * 2
-FROM 
+FROM
 (
-	SELECT DISTINCT ON (tenant_id) tenant_id, max(dept) as max_dept FROM 
+	SELECT DISTINCT ON (tenant_id) tenant_id, max(dept) as max_dept FROM
 	(
-		SELECT 
+		SELECT
 			second_distributed_table.dept, second_distributed_table.tenant_id
-		FROM 
+		FROM
 			second_distributed_table, distributed_table
-		WHERE 
+		WHERE
 			distributed_table.tenant_id = second_distributed_table.tenant_id
 	) foo_inner
 	GROUP BY
-		tenant_id 
+		tenant_id
 	ORDER BY 1 DESC
 ) as foo
 WHERE
-	foo.tenant_id != second_distributed_table.tenant_id 
+	foo.tenant_id != second_distributed_table.tenant_id
 	AND second_distributed_table.dept IN (2)
 RETURNING
 	second_distributed_table.tenant_id, second_distributed_table.dept;
@@ -103,29 +103,29 @@ DEBUG:  Plan 6 query after replacing subqueries and CTEs: UPDATE recursive_dml_q
 
 -- the subquery foo is recursively planned
 -- and foo itself is a non colocated subquery and recursively planned
-UPDATE 
-	second_distributed_table 
-SET 
+UPDATE
+	second_distributed_table
+SET
 	dept = foo.tenant_id::int / 4
-FROM 
+FROM
 (
-	SELECT DISTINCT foo_inner_1.tenant_id FROM 
+	SELECT DISTINCT foo_inner_1.tenant_id FROM
 	(
-		SELECT 
+		SELECT
 			second_distributed_table.dept, second_distributed_table.tenant_id
-		FROM 
+		FROM
 			second_distributed_table, distributed_table
-		WHERE 
+		WHERE
 			distributed_table.tenant_id = second_distributed_table.tenant_id
 		AND
 			second_distributed_table.dept IN (3,4)
 	) foo_inner_1,
 	(
-		SELECT 
-			second_distributed_table.tenant_id 
-		FROM 
+		SELECT
+			second_distributed_table.tenant_id
+		FROM
 			second_distributed_table, distributed_table
-		WHERE 
+		WHERE
 			distributed_table.tenant_id = second_distributed_table.tenant_id
 		AND
 			second_distributed_table.dept IN (4,5)
@@ -133,21 +133,21 @@ FROM
 	WHERE foo_inner_1.tenant_id != foo_inner_2.tenant_id
 ) as foo
 WHERE
-	foo.tenant_id != second_distributed_table.tenant_id 
+	foo.tenant_id != second_distributed_table.tenant_id
 	AND second_distributed_table.dept IN (3);
 DEBUG:  generating subplan 8_1 for subquery SELECT second_distributed_table.tenant_id FROM recursive_dml_queries.second_distributed_table, recursive_dml_queries.distributed_table WHERE ((distributed_table.tenant_id OPERATOR(pg_catalog.=) second_distributed_table.tenant_id) AND (second_distributed_table.dept OPERATOR(pg_catalog.=) ANY (ARRAY[4, 5])))
 DEBUG:  generating subplan 8_2 for subquery SELECT DISTINCT foo_inner_1.tenant_id FROM (SELECT second_distributed_table.dept, second_distributed_table.tenant_id FROM recursive_dml_queries.second_distributed_table, recursive_dml_queries.distributed_table WHERE ((distributed_table.tenant_id OPERATOR(pg_catalog.=) second_distributed_table.tenant_id) AND (second_distributed_table.dept OPERATOR(pg_catalog.=) ANY (ARRAY[3, 4])))) foo_inner_1, (SELECT intermediate_result.tenant_id FROM read_intermediate_result('8_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) foo_inner_2 WHERE (foo_inner_1.tenant_id OPERATOR(pg_catalog.<>) foo_inner_2.tenant_id)
 DEBUG:  Plan 8 query after replacing subqueries and CTEs: UPDATE recursive_dml_queries.second_distributed_table SET dept = ((foo.tenant_id)::integer OPERATOR(pg_catalog./) 4) FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('8_2'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) foo WHERE ((foo.tenant_id OPERATOR(pg_catalog.<>) second_distributed_table.tenant_id) AND (second_distributed_table.dept OPERATOR(pg_catalog.=) 3))
 -- we currently do not allow local tables in modification queries
-UPDATE 
-	distributed_table 
-SET 
+UPDATE
+	distributed_table
+SET
 	dept = avg_tenant_id::int
-FROM 
+FROM
 (
-	SELECT 
+	SELECT
 		avg(local_table.id::int) as avg_tenant_id
-	FROM 
+	FROM
 		local_table
 ) as foo
 WHERE
@@ -162,15 +162,15 @@ DEBUG:  Plan 11 query after replacing subqueries and CTEs: UPDATE recursive_dml_
 (1 row)
 
 -- we currently do not allow views in modification queries
-UPDATE 
-	distributed_table 
-SET 
+UPDATE
+	distributed_table
+SET
 	dept = avg_tenant_id::int
-FROM 
+FROM
 (
-	SELECT 
+	SELECT
 		avg(tenant_id::int) as avg_tenant_id
-	FROM 
+	FROM
 		tenant_ids
 ) as foo
 WHERE
@@ -184,32 +184,32 @@ DEBUG:  Plan 12 query after replacing subqueries and CTEs: UPDATE recursive_dml_
  50        |   50 | {"f1": 50, "f2": 2500}
 (1 row)
 
--- there is a lateral join (e.g., corrolated subquery) thus the subqueries cannot be 
+-- there is a lateral join (e.g., corrolated subquery) thus the subqueries cannot be
 -- recursively planned
-UPDATE 
-	second_distributed_table 
-SET 
+UPDATE
+	second_distributed_table
+SET
 	dept = foo.tenant_id::int / 4
-FROM 
+FROM
 (
-	SELECT DISTINCT foo_inner_1.tenant_id FROM 
+	SELECT DISTINCT foo_inner_1.tenant_id FROM
 	(
-		SELECT 
+		SELECT
 			second_distributed_table.dept, second_distributed_table.tenant_id
-		FROM 
+		FROM
 			second_distributed_table, distributed_table
-		WHERE 
+		WHERE
 			distributed_table.tenant_id = second_distributed_table.tenant_id
 		AND
 			second_distributed_table.dept IN (3,4)
-	) 
+	)
 	foo_inner_1 JOIN LATERAL
 	(
-		SELECT 
-			second_distributed_table.tenant_id 
-		FROM 
+		SELECT
+			second_distributed_table.tenant_id
+		FROM
 			second_distributed_table, distributed_table
-		WHERE 
+		WHERE
 			distributed_table.tenant_id = second_distributed_table.tenant_id
 			AND foo_inner_1.dept = second_distributed_table.dept
 		AND
@@ -222,55 +222,54 @@ ERROR:  complex joins are only supported when all distributed tables are joined
 -- again a corrolated subquery
 -- this time distribution key eq. exists
 -- however recursive planning is prevented due to correlated subqueries
-UPDATE 
-	second_distributed_table 
-SET 
+UPDATE
+	second_distributed_table
+SET
 	dept = foo.tenant_id::int / 4
-FROM 
+FROM
 (
-	SELECT baz.tenant_id FROM 
+	SELECT baz.tenant_id FROM
 	(
-		SELECT 
+		SELECT
 			second_distributed_table.dept, second_distributed_table.tenant_id
-		FROM 
+		FROM
 			second_distributed_table, distributed_table as d1
-		WHERE 
+		WHERE
 			d1.tenant_id = second_distributed_table.tenant_id
 		AND
 			second_distributed_table.dept IN (3,4)
-			AND 
-			second_distributed_table.tenant_id IN 
+			AND
+			second_distributed_table.tenant_id IN
 			(
 					SELECT s2.tenant_id
 					FROM second_distributed_table as s2
 					GROUP BY d1.tenant_id, s2.tenant_id
-			) 
+			)
 	) as baz
 	) as foo WHERE second_distributed_table.tenant_id = foo.tenant_id
 RETURNING *;
 ERROR:  complex joins are only supported when all distributed tables are joined on their distribution columns with equal operator
 -- we don't support subqueries/CTEs inside VALUES
-INSERT INTO 
-	second_distributed_table (tenant_id, dept) 
+INSERT INTO
+	second_distributed_table (tenant_id, dept)
 VALUES ('3', (WITH  vals AS (SELECT 3) select * from vals));
 DEBUG:  generating subplan 18_1 for CTE vals: SELECT 3
 DEBUG:  Plan 18 query after replacing subqueries and CTEs: INSERT INTO recursive_dml_queries.second_distributed_table (tenant_id, dept) VALUES ('3'::text, (SELECT vals."?column?" FROM (SELECT intermediate_result."?column?" FROM read_intermediate_result('18_1'::text, 'binary'::citus_copy_format) intermediate_result("?column?" integer)) vals))
 ERROR:  subqueries are not supported within INSERT queries
 HINT:  Try rewriting your queries with 'INSERT INTO ... SELECT' syntax.
-INSERT INTO 
-	second_distributed_table (tenant_id, dept) 
+INSERT INTO
+	second_distributed_table (tenant_id, dept)
 VALUES ('3', (SELECT 3));
 ERROR:  subqueries are not supported within INSERT queries
 HINT:  Try rewriting your queries with 'INSERT INTO ... SELECT' syntax.
 -- DML with an unreferenced SELECT CTE
 WITH cte_1 AS (
     WITH cte_2 AS (
-        SELECT tenant_id as cte2_id 
-        FROM second_distributed_table 
+        SELECT tenant_id as cte2_id
+        FROM second_distributed_table
         WHERE dept >= 2
     )
-    
-    UPDATE distributed_table 
+    UPDATE distributed_table
     SET dept = 10
     RETURNING *
 )
@@ -282,12 +281,11 @@ DEBUG:  generating subplan 20_1 for CTE cte_1: WITH cte_2 AS (SELECT second_dist
 DEBUG:  Plan 20 query after replacing subqueries and CTEs: UPDATE recursive_dml_queries.distributed_table SET dept = 5 FROM (SELECT intermediate_result.tenant_id, intermediate_result.dept, intermediate_result.info FROM read_intermediate_result('20_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text, dept integer, info jsonb)) cte_1 WHERE (distributed_table.tenant_id OPERATOR(pg_catalog.<) cte_1.tenant_id)
 WITH cte_1 AS (
     WITH cte_2 AS (
-        SELECT tenant_id as cte2_id 
-        FROM second_distributed_table 
+        SELECT tenant_id as cte2_id
+        FROM second_distributed_table
         WHERE dept >= 2
     )
-    
-    UPDATE distributed_table 
+    UPDATE distributed_table
     SET dept = 10
     RETURNING *
 )
@@ -299,13 +297,13 @@ DEBUG:  generating subplan 22_1 for CTE cte_1: WITH cte_2 AS (SELECT second_dist
 DEBUG:  Plan 22 query after replacing subqueries and CTEs: UPDATE recursive_dml_queries.distributed_table SET dept = 5 FROM (SELECT intermediate_result.tenant_id, intermediate_result.dept, intermediate_result.info FROM read_intermediate_result('22_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text, dept integer, info jsonb)) cte_1 WHERE (distributed_table.tenant_id OPERATOR(pg_catalog.<) cte_1.tenant_id)
 -- we don't support updating local table with a join with
 -- distributed tables
-UPDATE 
-	local_table 
-SET 
+UPDATE
+	local_table
+SET
 	id = 'citus_test'
-FROM 
+FROM
 	distributed_table
-WHERE 
+WHERE
 	distributed_table.tenant_id = local_table.id;
 ERROR:  relation local_table is not distributed
 RESET client_min_messages;

src/test/regress/expected/failure_add_disable_node.out

@@ -4,7 +4,6 @@
 -- master_disable_node and master_add_inactive_node can not be
 -- tested as they don't create network activity
 --
- 
 SELECT citus.mitmproxy('conn.allow()');
  mitmproxy 
 -----------

src/test/regress/expected/failure_cte_subquery.out

@@ -87,7 +87,6 @@ ERROR:  connection error: localhost:9060
 DETAIL:  server closed the connection unexpectedly
 	This probably means the server terminated abnormally
 	before or while processing the request.
-	  
 -- kill at the third copy (pull)
 SELECT citus.mitmproxy('conn.onQuery(query="SELECT DISTINCT users_table.user").kill()');
  mitmproxy 

src/test/regress/expected/local_shard_execution.out

@@ -29,7 +29,7 @@ SELECT create_distributed_table('second_distributed_table','key');
 INSERT INTO reference_table VALUES (1);
 INSERT INTO distributed_table VALUES (1, '1', 20);
 INSERT INTO second_distributed_table VALUES (1, '1');
--- a simple test for 
+-- a simple test for
 CREATE TABLE collections_list (
 	key bigserial,
 	ser bigserial,
@@ -44,7 +44,7 @@ SELECT create_distributed_table('collections_list', 'key');
  
 (1 row)
 
-CREATE TABLE collections_list_0 
+CREATE TABLE collections_list_0
 	PARTITION OF collections_list (key, ser, ts, collection_id, value)
 	FOR VALUES IN ( 0 );
 -- connection worker and get ready for the tests
@@ -54,19 +54,19 @@ SET search_path TO local_shard_execution;
 -- on the distributed tables (e.g., WHERE key = 1), we'll hit a shard
 -- placement which is local to this not
 CREATE OR REPLACE FUNCTION shard_of_distribution_column_is_local(dist_key int) RETURNS bool AS $$
-	
+
 		DECLARE shard_is_local BOOLEAN := FALSE;
 
 		BEGIN
-		  
+
 		  	WITH  local_shard_ids 			  AS (SELECT get_shard_id_for_distribution_column('local_shard_execution.distributed_table', dist_key)),
 				  all_local_shard_ids_on_node AS (SELECT shardid FROM pg_dist_placement WHERE groupid IN (SELECT groupid FROM pg_dist_local_group))
-		SELECT 
+		SELECT
 			true INTO shard_is_local
-		FROM 
-			local_shard_ids 
-		WHERE 
-			get_shard_id_for_distribution_column IN (SELECT * FROM all_local_shard_ids_on_node); 
+		FROM
+			local_shard_ids
+		WHERE
+			get_shard_id_for_distribution_column IN (SELECT * FROM all_local_shard_ids_on_node);
 
 		IF shard_is_local IS NULL THEN
 			shard_is_local = FALSE;
@@ -76,7 +76,7 @@ CREATE OR REPLACE FUNCTION shard_of_distribution_column_is_local(dist_key int) R
         END;
 $$ LANGUAGE plpgsql;
 -- pick some example values that reside on the shards locally and remote
--- distribution key values of 1,6, 500 and 701 are LOCAL to shards, 
+-- distribution key values of 1,6, 500 and 701 are LOCAL to shards,
 -- we'll use these values in the tests
 SELECT shard_of_distribution_column_is_local(1);
  shard_of_distribution_column_is_local 
@@ -102,7 +102,7 @@ SELECT shard_of_distribution_column_is_local(701);
  t
 (1 row)
 
--- distribution key values of 11 and 12 are REMOTE to shards 
+-- distribution key values of 11 and 12 are REMOTE to shards
 SELECT shard_of_distribution_column_is_local(11);
  shard_of_distribution_column_is_local 
 ---------------------------------------
@@ -118,7 +118,7 @@ SELECT shard_of_distribution_column_is_local(12);
 --- enable logging to see which tasks are executed locally
 SET client_min_messages TO LOG;
 SET citus.log_local_commands TO ON;
--- first, make sure that local execution works fine 
+-- first, make sure that local execution works fine
 -- with simple queries that are not in transcation blocks
 SELECT count(*) FROM distributed_table WHERE key = 1;
 LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_execution.distributed_table_1470001 distributed_table WHERE (key OPERATOR(pg_catalog.=) 1)
@@ -167,15 +167,15 @@ DELETE FROM second_distributed_table;
 -- load some more data for the following tests
 INSERT INTO second_distributed_table VALUES (1, '1');
 LOG:  executing the command locally: INSERT INTO local_shard_execution.second_distributed_table_1470005 (key, value) VALUES (1, '1'::text)
--- INSERT .. SELECT hitting a single single (co-located) shard(s) should 
+-- INSERT .. SELECT hitting a single single (co-located) shard(s) should
 -- be executed locally
-INSERT INTO distributed_table 
-SELECT 
-	distributed_table.* 
-FROM 
-	distributed_table, second_distributed_table 
-WHERE 
-	distributed_table.key = 1 and distributed_table.key=second_distributed_table.key 
+INSERT INTO distributed_table
+SELECT
+	distributed_table.*
+FROM
+	distributed_table, second_distributed_table
+WHERE
+	distributed_table.key = 1 and distributed_table.key=second_distributed_table.key
 ON CONFLICT(key) DO UPDATE SET value = '22'
 RETURNING *;
 LOG:  executing the command locally: INSERT INTO local_shard_execution.distributed_table_1470001 AS citus_table_alias (key, value, age) SELECT distributed_table.key, distributed_table.value, distributed_table.age FROM local_shard_execution.distributed_table_1470001 distributed_table, local_shard_execution.second_distributed_table_1470005 second_distributed_table WHERE (((distributed_table.key OPERATOR(pg_catalog.=) 1) AND (distributed_table.key OPERATOR(pg_catalog.=) second_distributed_table.key)) AND ((worker_hash(distributed_table.key) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(distributed_table.key) OPERATOR(pg_catalog.<=) '-1073741825'::integer))) ON CONFLICT(key) DO UPDATE SET value = '22'::text RETURNING citus_table_alias.key, citus_table_alias.value, citus_table_alias.age
@@ -185,13 +185,13 @@ LOG:  executing the command locally: INSERT INTO local_shard_execution.distribut
 (1 row)
 
 -- INSERT .. SELECT hitting multi-shards should go thourgh distributed execution
-INSERT INTO distributed_table 
-SELECT 
-	distributed_table.* 
-FROM 
-	distributed_table, second_distributed_table 
-WHERE 
-	distributed_table.key != 1 and distributed_table.key=second_distributed_table.key 
+INSERT INTO distributed_table
+SELECT
+	distributed_table.*
+FROM
+	distributed_table, second_distributed_table
+WHERE
+	distributed_table.key != 1 and distributed_table.key=second_distributed_table.key
 ON CONFLICT(key) DO UPDATE SET value = '22'
 RETURNING *;
  key | value | age 
@@ -251,11 +251,11 @@ LOG:  executing the command locally: SELECT key, value, age FROM local_shard_exe
 COPY reference_table FROM STDIN;
 COPY distributed_table FROM STDIN WITH CSV;
 COPY second_distributed_table FROM STDIN WITH CSV;
--- the behaviour in transaction blocks is the following: 
+-- the behaviour in transaction blocks is the following:
 	-- (a) Unless the first query is a local query, always use distributed execution.
-	-- (b) If the executor has used local execution, it has to use local execution 
-	--     for the remaining of the transaction block. If that's not possible, the 
-	-- 	   executor has to error out (e.g., TRUNCATE is a utility command and we 
+	-- (b) If the executor has used local execution, it has to use local execution
+	--     for the remaining of the transaction block. If that's not possible, the
+	-- 	   executor has to error out (e.g., TRUNCATE is a utility command and we
 	--	   currently do not support local execution of utility commands)
 -- rollback should be able to rollback local execution
 BEGIN;
@@ -322,14 +322,14 @@ SELECT count(*) FROM second_distributed_table;
 -- that has done before
 BEGIN;
 	-- INSERT is executed locally
-	INSERT INTO distributed_table VALUES (1, '11',21) ON CONFLICT(key) DO UPDATE SET value = '23' RETURNING *; 
+	INSERT INTO distributed_table VALUES (1, '11',21) ON CONFLICT(key) DO UPDATE SET value = '23' RETURNING *;
 LOG:  executing the command locally: INSERT INTO local_shard_execution.distributed_table_1470001 AS citus_table_alias (key, value, age) VALUES (1, '11'::text, 21) ON CONFLICT(key) DO UPDATE SET value = '23'::text RETURNING citus_table_alias.key, citus_table_alias.value, citus_table_alias.age
  key | value | age 
 -----+-------+-----
    1 | 23    |  20
 (1 row)
 
-	-- since the INSERT is executed locally, the SELECT should also be 
+	-- since the INSERT is executed locally, the SELECT should also be
 	-- executed locally and see the changes
 	SELECT * FROM distributed_table WHERE key = 1 ORDER BY 1,2,3;
 LOG:  executing the command locally: SELECT key, value, age FROM local_shard_execution.distributed_table_1470001 distributed_table WHERE (key OPERATOR(pg_catalog.=) 1) ORDER BY key, value, age
@@ -370,7 +370,7 @@ LOG:  executing the command locally: SELECT key, value, age FROM local_shard_exe
 (0 rows)
 
 -- if we start with a distributed execution, we should keep
--- using that and never switch back to local execution 
+-- using that and never switch back to local execution
 BEGIN;
 	DELETE FROM distributed_table WHERE value = '11';
 	-- although this command could have been executed
@@ -517,8 +517,7 @@ LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_e
      0
 (1 row)
 
-	
-	INSERT INTO distributed_table (key) SELECT i FROM generate_series(1,10)i; 
+	INSERT INTO distributed_table (key) SELECT i FROM generate_series(1,10)i;
 ERROR:  cannot execute command because a local execution has already been done in the transaction
 DETAIL:  Some parallel commands cannot be executed if a previous command has already been executed locally
 HINT:  Try re-running the transaction with "SET LOCAL citus.enable_local_execution TO OFF;"
@@ -532,8 +531,7 @@ LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_e
      0
 (1 row)
 
-	
-	INSERT INTO distributed_table (key) SELECT key+1 FROM distributed_table; 
+	INSERT INTO distributed_table (key) SELECT key+1 FROM distributed_table;
 ERROR:  cannot execute command because a local execution has already been done in the transaction
 DETAIL:  Some parallel commands cannot be executed if a previous command has already been executed locally
 HINT:  Try re-running the transaction with "SET LOCAL citus.enable_local_execution TO OFF;"
@@ -584,7 +582,7 @@ CREATE OR REPLACE PROCEDURE only_local_execution() AS $$
 		BEGIN
 			INSERT INTO distributed_table VALUES (1, '11',21) ON CONFLICT(key) DO UPDATE SET value = '29';
 			SELECT count(*) INTO cnt FROM distributed_table WHERE key = 1;
-			DELETE FROM distributed_table WHERE key = 1;	
+			DELETE FROM distributed_table WHERE key = 1;
         END;
 $$ LANGUAGE plpgsql;
 CALL only_local_execution();
@@ -666,11 +664,11 @@ SELECT * FROM local_insert, distributed_local_mixed ORDER BY 1,2,3,4,5;
 
 -- router CTE pushdown
 WITH all_data AS (SELECT * FROM distributed_table WHERE key = 1)
-SELECT 
-	count(*) 
-FROM 
-	distributed_table, all_data 
-WHERE 
+SELECT
+	count(*)
+FROM
+	distributed_table, all_data
+WHERE
 	distributed_table.key = all_data.key AND distributed_table.key = 1;
 LOG:  executing the command locally: WITH all_data AS (SELECT distributed_table_1.key, distributed_table_1.value, distributed_table_1.age FROM local_shard_execution.distributed_table_1470001 distributed_table_1 WHERE (distributed_table_1.key OPERATOR(pg_catalog.=) 1)) SELECT count(*) AS count FROM local_shard_execution.distributed_table_1470001 distributed_table, all_data WHERE ((distributed_table.key OPERATOR(pg_catalog.=) all_data.key) AND (distributed_table.key OPERATOR(pg_catalog.=) 1))
  count 
@@ -684,13 +682,13 @@ INSERT INTO distributed_table VALUES (2, '29', 29);
 INSERT INTO second_distributed_table VALUES (2, '29');
 -- single shard that is not a local query followed by a local query
 WITH all_data AS (SELECT * FROM second_distributed_table WHERE key = 2)
-SELECT 
+SELECT
 	distributed_table.key
-FROM 
-	distributed_table, all_data 
-WHERE 
+FROM
+	distributed_table, all_data
+WHERE
 	distributed_table.value = all_data.value AND distributed_table.key = 1
-ORDER BY 
+ORDER BY
 	1 DESC;
  key 
 -----
@@ -699,15 +697,15 @@ ORDER BY
 
 -- multi-shard CTE is followed by a query which could be executed locally, but
 -- since the query started with a parallel query, it doesn't use local execution
--- note that if we allow Postgres to inline the CTE (e.g., not have the EXISTS 
--- subquery), then it'd pushdown the filters and the query becomes single-shard, 
+-- note that if we allow Postgres to inline the CTE (e.g., not have the EXISTS
+-- subquery), then it'd pushdown the filters and the query becomes single-shard,
 -- locally executable query
 WITH all_data AS (SELECT * FROM distributed_table)
-SELECT 
-	count(*) 
-FROM 
-	distributed_table, all_data 
-WHERE 
+SELECT
+	count(*)
+FROM
+	distributed_table, all_data
+WHERE
 	distributed_table.key = all_data.key AND distributed_table.key = 1
 	AND EXISTS (SELECT * FROM all_data);
  count 
@@ -719,11 +717,11 @@ WHERE
 -- a subquery that needs to be recursively planned and a parallel
 -- query, so do not use local execution
 WITH all_data AS (SELECT age FROM distributed_table)
-SELECT 
-	count(*) 
-FROM 
-	distributed_table, all_data 
-WHERE 
+SELECT
+	count(*)
+FROM
+	distributed_table, all_data
+WHERE
 	distributed_table.key = all_data.age AND distributed_table.key = 1;
  count 
 -------
@@ -754,7 +752,7 @@ LOG:  executing the command locally: INSERT INTO local_shard_execution.distribut
    5 | 55    |  22
 (2 rows)
 
--- distributed execution of multi-rows INSERTs, where some part of the execution 
+-- distributed execution of multi-rows INSERTs, where some part of the execution
 -- could have been done via local execution but the executor choose the other way around
 -- because the command is a multi-shard query
 INSERT INTO distributed_table VALUES (1, '11',21), (2,'22',22), (3,'33',33), (4,'44',44),(5,'55',55) ON CONFLICT(key) DO UPDATE SET value = (EXCLUDED.value::int + 1)::text RETURNING *;
@@ -866,8 +864,8 @@ LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_e
      4
 (1 row)
 
-COMMIT;	
--- failures of local execution should rollback both the 
+COMMIT;
+-- failures of local execution should rollback both the
 -- local execution and remote executions
 -- fail on a local execution
 BEGIN;
@@ -948,7 +946,6 @@ LOG:  executing the command locally: DELETE FROM local_shard_execution.distribut
 LOG:  executing the command locally: DELETE FROM local_shard_execution.distributed_table_1470003 distributed_table WHERE (value OPERATOR(pg_catalog.<>) '123123213123213'::text)
 ROLLBACK;
 BEGIN;
-	
 	DELETE FROM reference_table WHERE key = 500 RETURNING *;
 LOG:  executing the command locally: DELETE FROM local_shard_execution.reference_table_1470000 reference_table WHERE (key OPERATOR(pg_catalog.=) 500) RETURNING key
  key 
@@ -985,7 +982,7 @@ BEGIN;
 ROLLBACK;
 -- probably not a realistic case since views are not very
 -- well supported with MX
-CREATE VIEW v_local_query_execution AS 
+CREATE VIEW v_local_query_execution AS
 SELECT * FROM distributed_table WHERE key = 500;
 SELECT * FROM v_local_query_execution;
 LOG:  executing the command locally: SELECT key, value, age FROM (SELECT distributed_table.key, distributed_table.value, distributed_table.age FROM local_shard_execution.distributed_table_1470003 distributed_table WHERE (distributed_table.key OPERATOR(pg_catalog.=) 500)) v_local_query_execution
@@ -996,7 +993,7 @@ LOG:  executing the command locally: SELECT key, value, age FROM (SELECT distrib
 
 -- similar test, but this time the view itself is a non-local
 -- query, but the query on the view is local
-CREATE VIEW v_local_query_execution_2 AS 
+CREATE VIEW v_local_query_execution_2 AS
 SELECT * FROM distributed_table;
 SELECT * FROM v_local_query_execution_2 WHERE key = 500;
 LOG:  executing the command locally: SELECT key, value, age FROM (SELECT distributed_table.key, distributed_table.value, distributed_table.age FROM local_shard_execution.distributed_table_1470003 distributed_table) v_local_query_execution_2 WHERE (key OPERATOR(pg_catalog.=) 500)
@@ -1016,19 +1013,15 @@ BEGIN;
 (1 row)
 
     DELETE FROM distributed_table WHERE key = 500;
-	
     ROLLBACK TO SAVEPOINT my_savepoint;
-	
 	DELETE FROM distributed_table WHERE key = 500;
 COMMIT;
 -- even if we switch from local execution -> remote execution,
 -- we are able to use local execution after rollback
 BEGIN;
-	   
 	SAVEPOINT my_savepoint;
     DELETE FROM distributed_table WHERE key = 500;
 LOG:  executing the command locally: DELETE FROM local_shard_execution.distributed_table_1470003 distributed_table WHERE (key OPERATOR(pg_catalog.=) 500)
-	
 	SELECT count(*) FROM distributed_table;
 LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_execution.distributed_table_1470001 distributed_table WHERE true
 LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_execution.distributed_table_1470003 distributed_table WHERE true
@@ -1038,7 +1031,6 @@ LOG:  executing the command locally: SELECT count(*) AS count FROM local_shard_e
 (1 row)
 
     ROLLBACK TO SAVEPOINT my_savepoint;
-	
 	DELETE FROM distributed_table WHERE key = 500;
 LOG:  executing the command locally: DELETE FROM local_shard_execution.distributed_table_1470003 distributed_table WHERE (key OPERATOR(pg_catalog.=) 500)
 COMMIT;
@@ -1079,7 +1071,7 @@ LOG:  executing the command locally: SELECT key, ser, ts, collection_id, value F
 
 COMMIT;
 -- the final queries for the following CTEs are going to happen on the intermediate results only
--- one of them will be executed remotely, and the other is locally 
+-- one of them will be executed remotely, and the other is locally
 -- Citus currently doesn't allow using task_assignment_policy for intermediate results
 WITH distributed_local_mixed AS (INSERT INTO reference_table VALUES (1000) RETURNING *) SELECT * FROM distributed_local_mixed;
 LOG:  executing the command locally: INSERT INTO local_shard_execution.reference_table_1470000 (key) VALUES (1000) RETURNING key
@@ -1131,10 +1123,10 @@ LOG:  executing the command locally: DELETE FROM local_shard_execution.reference
 
 COMMIT;
 \c - - - :master_port
--- local execution with custom type 
+-- local execution with custom type
 SET citus.replication_model TO "streaming";
 SET citus.shard_replication_factor TO 1;
-CREATE TYPE invite_resp AS ENUM ('yes', 'no', 'maybe'); 
+CREATE TYPE invite_resp AS ENUM ('yes', 'no', 'maybe');
 CREATE TABLE event_responses (
   event_id int,
   user_id int,

src/test/regress/expected/multi_behavioral_analytics_basics.out

@@ -58,7 +58,7 @@ FROM (
       AND e.event_type IN (3, 4)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
       WHERE  e.user_id >= 1
@@ -127,7 +127,7 @@ SELECT
       users_table
     WHERE
       user_id >= 1 AND
-      user_id <= 3 AND    
+      user_id <= 3 AND
       users_table.value_1 > 3 AND users_table.value_1 < 5
     GROUP BY
       user_id
@@ -170,7 +170,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time
         FROM users_table
         WHERE
@@ -205,9 +204,8 @@ FROM users_table
 WHERE user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 1 AND value_1 <= 2)
     AND user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 3 AND value_1 <= 4)
     AND user_id IN (SELECT user_id FROM users_table WHERE  value_1 >= 5 AND value_1 <= 6);
-    
 -- get some statistics from the aggregated results to ensure the results are correct
-SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;    
+SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
  count | count |        avg         
 -------+-------+--------------------
      5 |     5 | 3.8000000000000000
@@ -300,7 +298,6 @@ SELECT user_id, value_2 FROM users_table WHERE
   value_2 >= 3
   AND  EXISTS (SELECT user_id FROM events_table WHERE event_type > 1 AND event_type <= 3 AND value_3 > 1 AND user_id=users_table.user_id)
   AND  NOT EXISTS (SELECT user_id FROM events_table WHERE event_type > 3 AND event_type <= 4  AND value_3 > 1 AND user_id=users_table.user_id);
-  
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
  count | count |        avg         
@@ -315,21 +312,20 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
 ------------------------------------
 TRUNCATE agg_results;
 INSERT INTO agg_results(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
   WHERE  value_1 > 1
-         AND value_1 < 3 
+         AND value_1 < 3
          AND value_2 >= 1
-         AND EXISTS (SELECT user_id 
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type > 1 
-                            AND event_type < 3 
+                     WHERE  event_type > 1
+                            AND event_type < 3
                             AND value_3 > 1
-                            AND user_id = users_table.user_id 
-                     GROUP  BY user_id 
+                            AND user_id = users_table.user_id
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
-                     
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
  count | count |        avg         
@@ -337,7 +333,6 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
      4 |     2 | 3.5000000000000000
 (1 row)
 
-                                  
 ------------------------------------
 ------------------------------------
 -- Find me all users_table who logged in more than once
@@ -348,7 +343,7 @@ INSERT INTO agg_results(user_id, value_1_agg)
 SELECT user_id, value_1 from
 (
   SELECT user_id, value_1 From users_table
-  WHERE value_2 > 1 and user_id = 1 GROUP BY value_1, user_id HAVING count(*) > 1 
+  WHERE value_2 > 1 and user_id = 1 GROUP BY value_1, user_id HAVING count(*) > 1
 ) as a;
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
@@ -372,8 +367,7 @@ And user_id in
   (select user_id
    From users_table
    Where value_1 = 2
-   And value_2 > 1);  
- 
+   And value_2 > 1);
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
  count | count |        avg         
@@ -398,23 +392,21 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
     34 |     6 | 3.4411764705882353
 (1 row)
 
-       
 ------------------------------------
 ------------------------------------
 -- Find me all the users_table who has done some event more than three times
 ------------------------------------
-------------------------------------    
+------------------------------------
 TRUNCATE agg_results;
 INSERT INTO agg_results(user_id)
 select user_id from
 (
-  select 
+  select
      user_id
-   from 
+   from
    	events_table
 where event_type = 4 group by user_id having count(*) > 3
 ) as a;
-   
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
  count | count |        avg         
@@ -422,7 +414,6 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
      4 |     4 | 2.5000000000000000
 (1 row)
 
-       
 ------------------------------------
 ------------------------------------
 -- Find my assets that have the highest probability and fetch their metadata
@@ -432,17 +423,17 @@ TRUNCATE agg_results;
 INSERT INTO agg_results(user_id, value_1_agg, value_3_agg)
 SELECT
     users_table.user_id, users_table.value_1, prob
-FROM 
+FROM
    users_table
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, (GREATEST(coalesce(ma.value_4, 0.0) / 250 + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
     	users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
     	short_list.user_id = ma.user_id and ma.value_1 < 3 and short_list.event_type < 3
-    ) temp 
-  ON users_table.user_id = temp.user_id 
+    ) temp
+  ON users_table.user_id = temp.user_id
   WHERE users_table.value_1 < 3;
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
@@ -456,17 +447,17 @@ TRUNCATE agg_results;
 INSERT INTO agg_results(user_id)
 SELECT
     DISTINCT users_ids.user_id
-FROM 
+FROM
    (SELECT DISTINCT user_id FROM users_table) as users_ids
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, ma.value_1, (GREATEST(coalesce(ma.value_4 / 250, 0.0) + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
     	users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
     	short_list.user_id = ma.user_id and ma.value_1 < 3 and short_list.event_type < 2
-    ) temp 
-  ON users_ids.user_id = temp.user_id 
+    ) temp
+  ON users_ids.user_id = temp.user_id
   WHERE temp.value_1 < 3;
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results;
@@ -480,17 +471,17 @@ TRUNCATE agg_results;
 INSERT INTO agg_results(user_id, value_1_agg, value_2_agg)
 SELECT
     DISTINCT ON (users_ids.user_id) users_ids.user_id, temp.value_1, prob
-FROM 
+FROM
    (SELECT DISTINCT user_id FROM users_table) as users_ids
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, ma.value_1, (GREATEST(coalesce(ma.value_4 / 250, 0.0) + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
       users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
       short_list.user_id = ma.user_id and ma.value_1 < 3 and short_list.event_type < 2
-    ) temp 
-  ON users_ids.user_id = temp.user_id 
+    ) temp
+  ON users_ids.user_id = temp.user_id
   WHERE temp.value_1 < 3
   ORDER BY 1, 2;
 SELECT count(*), count(DISTINCT user_id), avg(user_id), avg(value_1_agg) FROM agg_results;
@@ -504,16 +495,16 @@ TRUNCATE agg_results;
 INSERT INTO agg_results(user_id, value_1_agg, value_2_agg)
 SELECT
     DISTINCT ON (users_ids.user_id) users_ids.user_id, temp.value_1, prob
-FROM 
+FROM
    (SELECT DISTINCT ON (user_id) user_id, value_2 FROM users_table ORDER BY 1,2) as users_ids
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, ma.value_1, (GREATEST(coalesce(ma.value_4 / 250, 0.0) + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
       users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
       short_list.user_id = ma.user_id and ma.value_1 < 10 and short_list.event_type < 2
-    ) temp 
+    ) temp
   ON users_ids.user_id = temp.user_id
   ORDER BY 1, 2;
 SELECT count(*), count(DISTINCT user_id), avg(user_id), avg(value_1_agg) FROM agg_results;

src/test/regress/expected/multi_behavioral_analytics_single_shard_queries.out

@@ -41,8 +41,8 @@ FROM (
     SELECT u.user_id, e.event_type::text AS event, e.time
     FROM users_table AS u,
          events_table AS e
-    WHERE u.user_id = e.user_id AND 
-    (u.user_id = 1 OR u.user_id = 2) AND 
+    WHERE u.user_id = e.user_id AND
+    (u.user_id = 1 OR u.user_id = 2) AND
     (e.user_id = 1 OR e.user_id = 2)
       AND e.event_type IN (1, 2)
   ) t
@@ -90,10 +90,9 @@ FROM (
       AND e.event_type IN (3, 4)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
-      
       WHERE  e.user_id >= 1
       AND e.user_id <= 2
       AND e.event_type IN (5, 6)
@@ -133,11 +132,10 @@ FROM (
       AND e.event_type IN (3, 4)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
-      
-      WHERE 
+      WHERE
       (e.user_id = 2 OR e.user_id = 3)
       AND e.event_type IN (4, 5)
   ) t2 ON (t1.user_id = t2.user_id)
@@ -170,7 +168,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time
         FROM users_table
         WHERE
@@ -213,7 +210,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time
         FROM users_table
         WHERE
@@ -251,9 +247,8 @@ WHERE user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 1 AND value_1
     AND user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 3 AND value_1 <= 4)
     AND user_id IN (SELECT user_id FROM users_table WHERE  value_1 >= 5 AND value_1 <= 6)
     AND user_id = 1;
-    
 -- get some statistics from the aggregated results to ensure the results are correct
-SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;  
+SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |          avg           
 -------+-------+------------------------
      1 |     1 | 1.00000000000000000000
@@ -272,9 +267,8 @@ WHERE user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 1 AND value_1
     AND user_id IN (SELECT user_id FROM users_table WHERE value_1 >= 3 AND value_1 <= 4 AND (user_id = 1 OR user_id = 2))
     AND user_id IN (SELECT user_id FROM users_table WHERE  value_1 >= 5 AND value_1 <= 6 AND (user_id = 1 OR user_id = 2))
     AND (user_id = 1 OR user_id = 2);
-    
 -- get some statistics from the aggregated results to ensure the results are correct
-SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;  
+SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |          avg           
 -------+-------+------------------------
      1 |     1 | 1.00000000000000000000
@@ -330,7 +324,6 @@ SELECT user_id, value_2 FROM users_table WHERE
   AND user_id = 1
   AND  EXISTS (SELECT user_id FROM events_table WHERE event_type > 1 AND event_type <= 3 AND value_3 > 1 AND user_id=users_table.user_id)
   AND  NOT EXISTS (SELECT user_id FROM events_table WHERE event_type > 4 AND event_type <= 5 AND value_3 > 4 AND user_id=users_table.user_id);
-  
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |          avg           
@@ -350,7 +343,6 @@ SELECT user_id, value_2 FROM users_table WHERE
   AND (user_id = 1 OR user_id = 2)
   AND  EXISTS (SELECT user_id FROM events_table WHERE event_type > 1 AND event_type <= 3 AND value_3 > 1 AND user_id=users_table.user_id AND (user_id = 1 OR user_id = 2))
   AND  NOT EXISTS (SELECT user_id FROM events_table WHERE event_type > 4 AND event_type <= 5 AND value_3 > 4 AND user_id=users_table.user_id AND (user_id = 1 OR user_id = 2));
-  
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |        avg         
@@ -365,23 +357,22 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
 ------------------------------------
 TRUNCATE agg_results_second;
 INSERT INTO agg_results_second(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
   WHERE  value_1 > 1
          AND value_1 < 3
-         AND value_2 >= 1 
+         AND value_2 >= 1
          AND user_id = 3
-         AND EXISTS (SELECT user_id 
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type > 1 
-                            AND event_type < 3 
+                     WHERE  event_type > 1
+                            AND event_type < 3
                             AND value_3 > 1
                             AND user_id = users_table.user_id
                             AND user_id = 3
-                     GROUP  BY user_id 
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
-                     
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |        avg         
@@ -396,22 +387,21 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
 ------------------------------------
 TRUNCATE agg_results_second;
 INSERT INTO agg_results_second(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
   WHERE  value_1 > 1
          AND value_1 < 3
          AND value_2 >= 1
          AND (user_id = 3 or user_id = 4)
-         AND EXISTS (SELECT user_id 
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type = 2 
-                            AND value_3 > 1 
+                     WHERE  event_type = 2
+                            AND value_3 > 1
                             AND user_id = users_table.user_id
                             AND (user_id = 3 or user_id = 4)
-                     GROUP  BY user_id 
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
-                     
 -- get some statistics from the aggregated results to ensure the results are correct
 SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
  count | count |        avg         
@@ -419,4 +409,3 @@ SELECT count(*), count(DISTINCT user_id), avg(user_id) FROM agg_results_second;
      4 |     2 | 3.5000000000000000
 (1 row)
 
-                                  

src/test/regress/expected/multi_citus_tools.out

@@ -7,9 +7,9 @@ SET citus.next_shard_id TO 1240000;
 -- test with invalid port, prevent OS dependent warning from being displayed
 SET client_min_messages to ERROR;
 -- PG 9.5 does not show context for plpgsql raise
--- message whereas PG 9.6 shows. disabling it 
+-- message whereas PG 9.6 shows. disabling it
 -- for this test only to have consistent behavior
--- b/w PG 9.6+ and PG 9.5. 
+-- b/w PG 9.6+ and PG 9.5.
 \set SHOW_CONTEXT never
 SELECT * FROM master_run_on_worker(ARRAY['localhost']::text[], ARRAY['666']::int[],
 								   ARRAY['select count(*) from pg_dist_shard']::text[],
@@ -61,7 +61,6 @@ SELECT * FROM master_run_on_worker(ARRAY[:node_name]::text[], ARRAY[:node_port]:
  localhost |     57637 | f       | expected a single row in query result
 (1 row)
 
-								   
 -- send multiple queries
 SELECT * FROM master_run_on_worker(ARRAY[:node_name, :node_name]::text[],
 								   ARRAY[:node_port, :node_port]::int[],
@@ -185,7 +184,6 @@ SELECT * FROM master_run_on_worker(ARRAY[:node_name]::text[], ARRAY[:node_port]:
  localhost |     57637 | t       | DROP TABLE
 (1 row)
 
-	
 -- verify table is dropped
 SELECT * FROM master_run_on_worker(ARRAY[:node_name]::text[], ARRAY[:node_port]::int[],
 								   ARRAY['select count(*) from second_table']::text[],
@@ -226,7 +224,6 @@ SELECT * FROM master_run_on_worker(ARRAY[:node_name]::text[], ARRAY[:node_port]:
  localhost |     57637 | f       | expected a single row in query result
 (1 row)
 
-								   
 -- send multiple queries
 SELECT * FROM master_run_on_worker(ARRAY[:node_name, :node_name]::text[],
 								   ARRAY[:node_port, :node_port]::int[],

src/test/regress/expected/multi_create_shards.out

@@ -223,9 +223,7 @@ SELECT shardmaxvalue::integer - shardminvalue::integer AS shard_size
 DELETE FROM pg_dist_shard_placement
 	WHERE shardid IN (SELECT shardid FROM pg_dist_shard
 					   WHERE logicalrelid = 'foreign_table_to_distribute'::regclass);
-	
 DELETE FROM pg_dist_shard
 	WHERE logicalrelid = 'foreign_table_to_distribute'::regclass;
-	
 DELETE FROM pg_dist_partition
 	WHERE logicalrelid = 'foreign_table_to_distribute'::regclass;

src/test/regress/expected/multi_distributed_transaction_id.out

@@ -1,6 +1,6 @@
 --
 -- MULTI_DISTRIBUTED_TRANSACTION_ID
--- 
+--
 -- Unit tests for distributed transaction id functionality
 --
 -- get the current transaction id, which should be uninitialized
@@ -17,7 +17,6 @@ SELECT initiator_node_identifier, transaction_number, transaction_stamp FROM get
 (1 row)
 
 BEGIN;
-	
 	-- we should still see the uninitialized values
 	SELECT initiator_node_identifier, transaction_number, transaction_stamp, (process_id = pg_backend_pid()) FROM get_current_transaction_id();
  initiator_node_identifier | transaction_number | transaction_stamp | ?column? 
@@ -52,7 +51,6 @@ SELECT initiator_node_identifier, transaction_number, transaction_stamp, (proces
 
 -- also see that ROLLBACK (i.e., failures in the transaction) clears the shared memory
 BEGIN;
-	
 	-- we should still see the uninitialized values
 	SELECT initiator_node_identifier, transaction_number, transaction_stamp, (process_id = pg_backend_pid()) FROM get_current_transaction_id();
  initiator_node_identifier | transaction_number | transaction_stamp | ?column? 
@@ -79,7 +77,6 @@ COMMIT;
 
 -- we should also see that a new connection means an uninitialized transaction id
 BEGIN;
-	
 	SELECT assign_distributed_transaction_id(52, 52, '2015-01-01 00:00:00+0');
  assign_distributed_transaction_id 
 -----------------------------------
@@ -126,7 +123,7 @@ ROLLBACK PREPARED 'dist_xact_id_test';
 -- set back to the original zone
 SET TIME ZONE DEFAULT;
 -- parallel safe wrapper for getting the current transaction number
-CREATE OR REPLACE FUNCTION parallel_worker_transaction_id_test() 
+CREATE OR REPLACE FUNCTION parallel_worker_transaction_id_test()
 RETURNS bigint STRICT VOLATILE PARALLEL SAFE AS $$
 	SELECT transaction_number FROM get_current_transaction_id();
 $$ LANGUAGE sql;

src/test/regress/expected/multi_foreign_key_relation_graph.out

@@ -161,7 +161,7 @@ SELECT get_referencing_relation_id_list::regclass FROM get_referencing_relation_
 ----------------------------------
 (0 rows)
 
--- some tests within transction blocks to make sure that 
+-- some tests within transction blocks to make sure that
 -- cache invalidation works fine
 CREATE TABLE test_1 (id int UNIQUE);
 CREATE TABLE test_2 (id int UNIQUE);
@@ -198,17 +198,17 @@ SELECT create_distributed_Table('test_5', 'id');
  
 (1 row)
 
-CREATE VIEW referential_integrity_summary AS 
-    WITH RECURSIVE referential_integrity_summary(n, table_name, referencing_relations, referenced_relations) AS 
+CREATE VIEW referential_integrity_summary AS
+    WITH RECURSIVE referential_integrity_summary(n, table_name, referencing_relations, referenced_relations) AS
     (
         SELECT 0,'0','{}'::regclass[],'{}'::regclass[]
       UNION ALL
-        SELECT 
-          n + 1, 
-          'test_' || n + 1|| '' as table_name, 
-          (SELECT  array_agg(get_referencing_relation_id_list::regclass ORDER BY 1) FROM get_referencing_relation_id_list(('test_' || (n +1) ) ::regclass)) as referencing_relations, 
+        SELECT
+          n + 1,
+          'test_' || n + 1|| '' as table_name,
+          (SELECT  array_agg(get_referencing_relation_id_list::regclass ORDER BY 1) FROM get_referencing_relation_id_list(('test_' || (n +1) ) ::regclass)) as referencing_relations,
           (SELECT  array_agg(get_referenced_relation_id_list::regclass ORDER BY 1) FROM get_referenced_relation_id_list(('test_' || (n +1) ) ::regclass)) as referenced_by_relations
-        FROM referential_integrity_summary, pg_class 
+        FROM referential_integrity_summary, pg_class
         WHERE
          pg_class.relname = ('test_' || (n +1))
         AND n < 5
@@ -216,7 +216,7 @@ CREATE VIEW referential_integrity_summary AS
     SELECT * FROM referential_integrity_summary WHERE n != 0 ORDER BY 1;
 -- make sure that invalidation through ALTER TABLE works fine
 BEGIN;
-    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);   
+    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);
     SELECT * FROM referential_integrity_summary;
  n | table_name | referencing_relations | referenced_relations 
 ---+------------+-----------------------+----------------------
@@ -263,7 +263,7 @@ BEGIN;
 ROLLBACK;
 -- similar test, but slightly different order of creating foreign keys
 BEGIN;
-    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);   
+    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);
     SELECT * FROM referential_integrity_summary;
  n | table_name | referencing_relations | referenced_relations 
 ---+------------+-----------------------+----------------------
@@ -310,7 +310,7 @@ BEGIN;
 ROLLBACK;
 -- make sure that DROP CONSTRAINT works invalidates the cache correctly
 BEGIN;
-    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);   
+    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);
     ALTER TABLE test_3 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_2(id);
     ALTER TABLE test_4 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_3(id);
     ALTER TABLE test_5 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_4(id);
@@ -412,7 +412,7 @@ COMMIT;
 -- DROP TABLE works expected
 -- re-create the constraints
 BEGIN;
-    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);   
+    ALTER TABLE test_2 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_1(id);
     ALTER TABLE test_3 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_2(id);
     ALTER TABLE test_4 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_3(id);
     ALTER TABLE test_5 ADD CONSTRAINT fkey_1 FOREIGN KEY(id) REFERENCES test_4(id);
@@ -601,7 +601,6 @@ drop cascades to table test_8
 ---------------------------------
 (0 rows)
 
-    
 ROLLBACK;
 SET search_path TO public;
 DROP SCHEMA fkey_graph CASCADE;

src/test/regress/expected/multi_insert_select.out

@@ -73,10 +73,10 @@ SET client_min_messages TO INFO;
 SELECT
    raw_events_first.user_id
 FROM
-   raw_events_first, raw_events_second 
+   raw_events_first, raw_events_second
 WHERE
    raw_events_first.user_id = raw_events_second.user_id
-ORDER BY 
+ORDER BY
   user_id DESC;
  user_id 
 ---------
@@ -161,10 +161,10 @@ INSERT INTO raw_events_first (user_id, time, value_1, value_2, value_3, value_4)
                          (8, now(), 80, 800, 8000, 80000);
 -- reorder columns
 SET client_min_messages TO DEBUG2;
-INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time) 
-SELECT 
-   value_2, value_1, value_3, value_4, user_id, time 
-FROM 
+INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time)
+SELECT
+   value_2, value_1, value_3, value_4, user_id, time
+FROM
    raw_events_first
 WHERE
    user_id = 8;
@@ -174,10 +174,10 @@ DEBUG:  Skipping target shard interval 13300006 since SELECT query for it pruned
 DEBUG:  Skipping target shard interval 13300007 since SELECT query for it pruned away
 DEBUG:  Plan is router executable
 -- a zero shard select
-INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time) 
-SELECT 
-   value_2, value_1, value_3, value_4, user_id, time 
-FROM 
+INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time)
+SELECT
+   value_2, value_1, value_3, value_4, user_id, time
+FROM
    raw_events_first
 WHERE
    false;
@@ -187,10 +187,10 @@ DEBUG:  Skipping target shard interval 13300006 since SELECT query for it pruned
 DEBUG:  Skipping target shard interval 13300007 since SELECT query for it pruned away
 DEBUG:  Plan is router executable
 -- another zero shard select
-INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time) 
-SELECT 
-   value_2, value_1, value_3, value_4, user_id, time 
-FROM 
+INSERT INTO raw_events_second (value_2, value_1, value_3, value_4, user_id, time)
+SELECT
+   value_2, value_1, value_3, value_4, user_id, time
+FROM
    raw_events_first
 WHERE
    0 != 0;
@@ -205,13 +205,13 @@ INSERT INTO raw_events_first (user_id, time, value_1, value_2, value_3, value_4)
                          (9, now(), 90, 900, 9000, 90000);
 -- show that RETURNING also works
 SET client_min_messages TO DEBUG2;
-INSERT INTO raw_events_second (user_id, value_1, value_3) 
-SELECT 
+INSERT INTO raw_events_second (user_id, value_1, value_3)
+SELECT
    user_id, value_1, value_3
 FROM
-   raw_events_first 
+   raw_events_first
 WHERE
-   value_3 = 9000 
+   value_3 = 9000
 RETURNING *;
 DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300004 AS citus_table_alias (user_id, value_1, value_3) SELECT user_id, value_1, value_3 FROM public.raw_events_first_13300000 raw_events_first WHERE ((value_3 OPERATOR(pg_catalog.=) (9000)::double precision) AND ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1073741825'::integer))) RETURNING citus_table_alias.user_id, citus_table_alias."time", citus_table_alias.value_1, citus_table_alias.value_2, citus_table_alias.value_3, citus_table_alias.value_4
 DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id, value_1, value_3) SELECT user_id, value_1, value_3 FROM public.raw_events_first_13300001 raw_events_first WHERE ((value_3 OPERATOR(pg_catalog.=) (9000)::double precision) AND ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-1073741824'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1'::integer))) RETURNING citus_table_alias.user_id, citus_table_alias."time", citus_table_alias.value_1, citus_table_alias.value_2, citus_table_alias.value_3, citus_table_alias.value_4
@@ -225,13 +225,13 @@ DEBUG:  Plan is router executable
 
 -- hits two shards
 \set VERBOSITY TERSE
-INSERT INTO raw_events_second (user_id, value_1, value_3) 
-SELECT 
+INSERT INTO raw_events_second (user_id, value_1, value_3)
+SELECT
    user_id, value_1, value_3
 FROM
-   raw_events_first 
+   raw_events_first
 WHERE
-   user_id = 9 OR user_id = 16 
+   user_id = 9 OR user_id = 16
 RETURNING *;
 DEBUG:  Skipping target shard interval 13300004 since SELECT query for it pruned away
 DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id, value_1, value_3) SELECT user_id, value_1, value_3 FROM public.raw_events_first_13300001 raw_events_first WHERE (((user_id OPERATOR(pg_catalog.=) 9) OR (user_id OPERATOR(pg_catalog.=) 16)) AND ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-1073741824'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1'::integer))) RETURNING citus_table_alias.user_id, citus_table_alias."time", citus_table_alias.value_1, citus_table_alias.value_2, citus_table_alias.value_3, citus_table_alias.value_4
@@ -240,9 +240,9 @@ DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300007 AS
 DEBUG:  Plan is router executable
 ERROR:  duplicate key value violates unique constraint "raw_events_second_user_id_value_1_key_13300007"
 -- now do some aggregations
-INSERT INTO agg_events 
+INSERT INTO agg_events
 SELECT
-   user_id, sum(value_1), avg(value_2), sum(value_3), count(value_4) 
+   user_id, sum(value_1), avg(value_2), sum(value_3), count(value_4)
 FROM
    raw_events_first
 GROUP BY
@@ -253,7 +253,7 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id, value_1_agg, value_2_agg, value_3_agg, value_4_agg) SELECT user_id, sum(value_1) AS sum, avg(value_2) AS avg, sum(value_3) AS sum, count(value_4) AS count FROM public.raw_events_first_13300003 raw_events_first WHERE ((worker_hash(user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) 2147483647)) GROUP BY user_id
 DEBUG:  Plan is router executable
 -- group by column not exists on the SELECT target list
-INSERT INTO agg_events (value_3_agg, value_4_agg, value_1_agg, user_id) 
+INSERT INTO agg_events (value_3_agg, value_4_agg, value_1_agg, user_id)
 SELECT
    sum(value_3), count(value_4), sum(value_1), user_id
 FROM
@@ -268,16 +268,16 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_t
 DEBUG:  Plan is router executable
 ERROR:  duplicate key value violates unique constraint "agg_events_user_id_value_1_agg_key_13300008"
 -- some subquery tests
-INSERT INTO agg_events 
-            (value_1_agg, 
-             user_id) 
-SELECT SUM(value_1), 
-       id 
-FROM   (SELECT raw_events_second.user_id AS id, 
-               raw_events_second.value_1 
-        FROM   raw_events_first, 
-               raw_events_second 
-        WHERE  raw_events_first.user_id = raw_events_second.user_id) AS foo 
+INSERT INTO agg_events
+            (value_1_agg,
+             user_id)
+SELECT SUM(value_1),
+       id
+FROM   (SELECT raw_events_second.user_id AS id,
+               raw_events_second.value_1
+        FROM   raw_events_first,
+               raw_events_second
+        WHERE  raw_events_first.user_id = raw_events_second.user_id) AS foo
 GROUP  BY id
 ORDER  BY id;
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300008 AS citus_table_alias (user_id, value_1_agg) SELECT id, sum(value_1) AS sum FROM (SELECT raw_events_second.user_id AS id, raw_events_second.value_1 FROM public.raw_events_first_13300000 raw_events_first, public.raw_events_second_13300004 raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id)) foo WHERE ((worker_hash(id) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(id) OPERATOR(pg_catalog.<=) '-1073741825'::integer)) GROUP BY id ORDER BY id
@@ -286,20 +286,20 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id, value_1_agg) SELECT id, sum(value_1) AS sum FROM (SELECT raw_events_second.user_id AS id, raw_events_second.value_1 FROM public.raw_events_first_13300003 raw_events_first, public.raw_events_second_13300007 raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id)) foo WHERE ((worker_hash(id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(id) OPERATOR(pg_catalog.<=) 2147483647)) GROUP BY id ORDER BY id
 DEBUG:  Plan is router executable
 ERROR:  duplicate key value violates unique constraint "agg_events_user_id_value_1_agg_key_13300008"
--- subquery one more level depth 
-INSERT INTO agg_events 
-            (value_4_agg, 
-             value_1_agg, 
-             user_id) 
-SELECT v4, 
-       v1, 
-       id 
-FROM   (SELECT SUM(raw_events_second.value_4) AS v4, 
-               SUM(raw_events_first.value_1) AS v1, 
-               raw_events_second.user_id      AS id 
-        FROM   raw_events_first, 
-               raw_events_second 
-        WHERE  raw_events_first.user_id = raw_events_second.user_id 
+-- subquery one more level depth
+INSERT INTO agg_events
+            (value_4_agg,
+             value_1_agg,
+             user_id)
+SELECT v4,
+       v1,
+       id
+FROM   (SELECT SUM(raw_events_second.value_4) AS v4,
+               SUM(raw_events_first.value_1) AS v1,
+               raw_events_second.user_id      AS id
+        FROM   raw_events_first,
+               raw_events_second
+        WHERE  raw_events_first.user_id = raw_events_second.user_id
         GROUP  BY raw_events_second.user_id) AS foo
 ORDER  BY id;
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300008 AS citus_table_alias (user_id, value_1_agg, value_4_agg) SELECT id, v1, v4 FROM (SELECT sum(raw_events_second.value_4) AS v4, sum(raw_events_first.value_1) AS v1, raw_events_second.user_id AS id FROM public.raw_events_first_13300000 raw_events_first, public.raw_events_second_13300004 raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.user_id) foo WHERE ((worker_hash(id) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(id) OPERATOR(pg_catalog.<=) '-1073741825'::integer)) ORDER BY id
@@ -451,19 +451,19 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT f2.id FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first_13300003 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) reference_table.user_id)) foo) f JOIN (SELECT foo2.v4, foo2.v1, foo2.id FROM (SELECT sum(raw_events_second.value_4) AS v4, sum(raw_events_first.value_1) AS v1, raw_events_second.user_id AS id FROM public.raw_events_first_13300003 raw_events_first, public.raw_events_second_13300007 raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) OPERATOR(pg_catalog.>) (1000)::numeric)) foo2) f2 ON ((f.id OPERATOR(pg_catalog.=) f2.id))) WHERE ((f.id OPERATOR(pg_catalog.=) ANY (SELECT raw_events_second.user_id FROM public.raw_events_second_13300007 raw_events_second)) AND ((worker_hash(f2.id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(f2.id) OPERATOR(pg_catalog.<=) 2147483647)))
 DEBUG:  Plan is router executable
 -- some UPSERTS
-INSERT INTO agg_events AS ae 
+INSERT INTO agg_events AS ae
             (
                         user_id,
                         value_1_agg,
                         agg_time
-            ) 
+            )
 SELECT user_id,
        value_1,
        time
 FROM   raw_events_first
 ON conflict (user_id, value_1_agg)
 DO UPDATE
-   SET    agg_time = EXCLUDED.agg_time 
+   SET    agg_time = EXCLUDED.agg_time
    WHERE  ae.agg_time < EXCLUDED.agg_time;
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300008 AS ae (user_id, value_1_agg, agg_time) SELECT user_id, value_1, "time" FROM public.raw_events_first_13300000 raw_events_first WHERE ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1073741825'::integer)) ON CONFLICT(user_id, value_1_agg) DO UPDATE SET agg_time = excluded.agg_time WHERE (ae.agg_time OPERATOR(pg_catalog.<) excluded.agg_time)
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS ae (user_id, value_1_agg, agg_time) SELECT user_id, value_1, "time" FROM public.raw_events_first_13300001 raw_events_first WHERE ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-1073741824'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1'::integer)) ON CONFLICT(user_id, value_1_agg) DO UPDATE SET agg_time = excluded.agg_time WHERE (ae.agg_time OPERATOR(pg_catalog.<) excluded.agg_time)
@@ -471,19 +471,19 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS ae (use
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS ae (user_id, value_1_agg, agg_time) SELECT user_id, value_1, "time" FROM public.raw_events_first_13300003 raw_events_first WHERE ((worker_hash(user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) 2147483647)) ON CONFLICT(user_id, value_1_agg) DO UPDATE SET agg_time = excluded.agg_time WHERE (ae.agg_time OPERATOR(pg_catalog.<) excluded.agg_time)
 DEBUG:  Plan is router executable
 -- upserts with returning
-INSERT INTO agg_events AS ae 
-            ( 
-                        user_id, 
-                        value_1_agg, 
-                        agg_time 
-            ) 
-SELECT user_id, 
-       value_1, 
-       time 
-FROM   raw_events_first 
+INSERT INTO agg_events AS ae
+            (
+                        user_id,
+                        value_1_agg,
+                        agg_time
+            )
+SELECT user_id,
+       value_1,
+       time
+FROM   raw_events_first
 ON conflict (user_id, value_1_agg)
 DO UPDATE
-   SET    agg_time = EXCLUDED.agg_time 
+   SET    agg_time = EXCLUDED.agg_time
    WHERE  ae.agg_time < EXCLUDED.agg_time
 RETURNING user_id, value_1_agg;
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300008 AS ae (user_id, value_1_agg, agg_time) SELECT user_id, value_1, "time" FROM public.raw_events_first_13300000 raw_events_first WHERE ((worker_hash(user_id) OPERATOR(pg_catalog.>=) '-2147483648'::integer) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) '-1073741825'::integer)) ON CONFLICT(user_id, value_1_agg) DO UPDATE SET agg_time = excluded.agg_time WHERE (ae.agg_time OPERATOR(pg_catalog.<) excluded.agg_time) RETURNING ae.user_id, ae.value_1_agg
@@ -584,10 +584,10 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_t
 DEBUG:  Plan is router executable
 SET client_min_messages TO INFO;
 -- see that the results are different from the SELECT query
-SELECT 
+SELECT
   user_id, value_1_agg
-FROM 
-  agg_events 
+FROM
+  agg_events
 ORDER BY
   user_id, value_1_agg;
  user_id | value_1_agg 
@@ -770,7 +770,6 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300002 raw_events_first LEFT JOIN public.raw_events_second_13300006 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 1073741823))
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300003 raw_events_first LEFT JOIN public.raw_events_second_13300007 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 2147483647))
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_second.user_id
@@ -781,7 +780,6 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id) SELECT raw_events_second.user_id FROM (public.reference_table_13300012 reference_table LEFT JOIN public.raw_events_second_13300006 raw_events_second ON ((reference_table.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((worker_hash(raw_events_second.user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(raw_events_second.user_id) OPERATOR(pg_catalog.<=) 1073741823))
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT raw_events_second.user_id FROM (public.reference_table_13300012 reference_table LEFT JOIN public.raw_events_second_13300007 raw_events_second ON ((reference_table.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((worker_hash(raw_events_second.user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(raw_events_second.user_id) OPERATOR(pg_catalog.<=) 2147483647))
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -793,7 +791,6 @@ DEBUG:  Skipping target shard interval 13300009 since SELECT query for it pruned
 DEBUG:  Skipping target shard interval 13300010 since SELECT query for it pruned away
 DEBUG:  Skipping target shard interval 13300011 since SELECT query for it pruned away
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -805,7 +802,6 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300002 raw_events_first LEFT JOIN (SELECT NULL::integer AS user_id, NULL::timestamp without time zone AS "time", NULL::integer AS value_1, NULL::integer AS value_2, NULL::double precision AS value_3, NULL::bigint AS value_4 WHERE false) raw_events_second(user_id, "time", value_1, value_2, value_3, value_4) ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE (((raw_events_second.user_id OPERATOR(pg_catalog.=) 10) OR (raw_events_second.user_id OPERATOR(pg_catalog.=) 11)) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 1073741823)))
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300003 raw_events_first LEFT JOIN public.raw_events_second_13300007 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE (((raw_events_second.user_id OPERATOR(pg_catalog.=) 10) OR (raw_events_second.user_id OPERATOR(pg_catalog.=) 11)) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 2147483647)))
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -817,7 +813,6 @@ DEBUG:  Skipping target shard interval 13300009 since SELECT query for it pruned
 DEBUG:  Skipping target shard interval 13300010 since SELECT query for it pruned away
 DEBUG:  Skipping target shard interval 13300011 since SELECT query for it pruned away
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -829,7 +824,6 @@ DEBUG:  Skipping target shard interval 13300009 since SELECT query for it pruned
 DEBUG:  Skipping target shard interval 13300010 since SELECT query for it pruned away
 DEBUG:  Skipping target shard interval 13300011 since SELECT query for it pruned away
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -841,7 +835,6 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300002 raw_events_first LEFT JOIN public.raw_events_second_13300006 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((raw_events_first.user_id OPERATOR(pg_catalog.=) ANY (ARRAY[19, 20, 21])) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 1073741823)))
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM ((SELECT NULL::integer AS user_id, NULL::timestamp without time zone AS "time", NULL::integer AS value_1, NULL::integer AS value_2, NULL::double precision AS value_3, NULL::bigint AS value_4 WHERE false) raw_events_first(user_id, "time", value_1, value_2, value_3, value_4) LEFT JOIN public.raw_events_second_13300007 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((raw_events_first.user_id OPERATOR(pg_catalog.=) ANY (ARRAY[19, 20, 21])) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 2147483647)))
 DEBUG:  Plan is router executable
- 
  INSERT INTO agg_events (user_id)
  SELECT
    raw_events_first.user_id
@@ -853,7 +846,6 @@ DEBUG:  distributed statement: INSERT INTO public.agg_events_13300009 AS citus_t
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300002 raw_events_first JOIN public.raw_events_second_13300006 raw_events_second ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((raw_events_second.user_id OPERATOR(pg_catalog.=) ANY (ARRAY[19, 20, 21])) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 1073741823)))
 DEBUG:  distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id) SELECT raw_events_first.user_id FROM (public.raw_events_first_13300003 raw_events_first JOIN (SELECT NULL::integer AS user_id, NULL::timestamp without time zone AS "time", NULL::integer AS value_1, NULL::integer AS value_2, NULL::double precision AS value_3, NULL::bigint AS value_4 WHERE false) raw_events_second(user_id, "time", value_1, value_2, value_3, value_4) ON ((raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id))) WHERE ((raw_events_second.user_id OPERATOR(pg_catalog.=) ANY (ARRAY[19, 20, 21])) AND ((worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(raw_events_first.user_id) OPERATOR(pg_catalog.<=) 2147483647)))
 DEBUG:  Plan is router executable
- 
  -- the following is a very tricky query for Citus
  -- although we do not support pushing down JOINs on non-partition
  -- columns here it is safe to push it down given that we're looking for
@@ -861,16 +853,15 @@ DEBUG:  Plan is router executable
  -- Note that the query always hits the same shard on raw_events_second
  -- and this query wouldn't have worked if we're to use different worker
  -- count or shard replication factor
- INSERT INTO agg_events 
-             (user_id) 
- SELECT raw_events_first.user_id 
- FROM   raw_events_first, 
-        raw_events_second 
- WHERE  raw_events_second.user_id = raw_events_first.value_1 
-        AND raw_events_first.value_1 = 12; 
+ INSERT INTO agg_events
+             (user_id)
+ SELECT raw_events_first.user_id
+ FROM   raw_events_first,
+        raw_events_second
+ WHERE  raw_events_second.user_id = raw_events_first.value_1
+        AND raw_events_first.value_1 = 12;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- some unsupported LEFT/INNER JOINs
  -- JOIN on one table with partition column other is not
  INSERT INTO agg_events (user_id)
@@ -880,7 +871,6 @@ DETAIL:  Select query cannot be pushed down to the worker.
    raw_events_first LEFT JOIN raw_events_second ON raw_events_first.user_id = raw_events_second.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- same as the above with INNER JOIN
  INSERT INTO agg_events (user_id)
  SELECT
@@ -889,17 +879,15 @@ DETAIL:  Select query cannot be pushed down to the worker.
    raw_events_first INNER JOIN raw_events_second ON raw_events_first.user_id = raw_events_second.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- a not meaningful query
- INSERT INTO agg_events 
-             (user_id) 
- SELECT raw_events_second.user_id 
- FROM   raw_events_first, 
-        raw_events_second 
- WHERE  raw_events_first.user_id = raw_events_first.value_1; 
+ INSERT INTO agg_events
+             (user_id)
+ SELECT raw_events_second.user_id
+ FROM   raw_events_first,
+        raw_events_second
+ WHERE  raw_events_first.user_id = raw_events_first.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- both tables joined on non-partition columns
  INSERT INTO agg_events (user_id)
  SELECT
@@ -908,7 +896,6 @@ DETAIL:  Select query cannot be pushed down to the worker.
    raw_events_first LEFT JOIN raw_events_second ON raw_events_first.value_1 = raw_events_second.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- same as the above with INNER JOIN
  INSERT INTO agg_events (user_id)
  SELECT
@@ -917,7 +904,6 @@ DETAIL:  Select query cannot be pushed down to the worker.
    raw_events_first INNER JOIN raw_events_second ON raw_events_first.value_1 = raw_events_second.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
 -- 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)
@@ -925,11 +911,10 @@ SELECT
   raw_events_first.user_id
 FROM
   raw_events_first LEFT JOIN raw_events_second ON raw_events_first.user_id = raw_events_second.value_1
-WHERE 
+WHERE
   raw_events_first.user_id = 10;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- same as the above with INNER JOIN
  INSERT INTO agg_events (user_id)
  SELECT
@@ -939,7 +924,6 @@ DETAIL:  Select query cannot be pushed down to the worker.
  WHERE raw_events_first.user_id = 10;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- make things a bit more complicate with IN clauses
  INSERT INTO agg_events (user_id)
  SELECT
@@ -949,32 +933,29 @@ DETAIL:  Select query cannot be pushed down to the worker.
    WHERE raw_events_first.value_1 IN (10, 11,12) OR raw_events_second.user_id IN (1,2,3,4);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- implicit join on non partition column should also not be pushed down
- INSERT INTO agg_events 
-             (user_id) 
- SELECT raw_events_first.user_id 
- FROM   raw_events_first, 
-        raw_events_second 
- WHERE  raw_events_second.user_id = raw_events_first.value_1; 
+ INSERT INTO agg_events
+             (user_id)
+ SELECT raw_events_first.user_id
+ FROM   raw_events_first,
+        raw_events_second
+ WHERE  raw_events_second.user_id = raw_events_first.value_1;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- the following is again a tricky query for Citus
  -- if the given filter was on value_1 as shown in the above, Citus could
  -- push it down. But here the query is refused
- INSERT INTO agg_events 
-             (user_id) 
- SELECT raw_events_first.user_id 
- FROM   raw_events_first, 
-        raw_events_second 
- WHERE  raw_events_second.user_id = raw_events_first.value_1 
+ INSERT INTO agg_events
+             (user_id)
+ SELECT raw_events_first.user_id
+ FROM   raw_events_first,
+        raw_events_second
+ WHERE  raw_events_second.user_id = raw_events_first.value_1
         AND raw_events_first.value_2 = 12;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- lets do some unsupported query tests with subqueries
- -- foo is not joined on the partition key so the query is not 
+ -- foo is not joined on the partition key so the query is not
  -- pushed down
  INSERT INTO agg_events
              (user_id, value_4_agg)
@@ -1008,17 +989,17 @@ ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
  -- if the given filter was on value_1 as shown in the above, Citus could
  -- push it down. But here the query is refused
- INSERT INTO agg_events 
-             (user_id) 
- SELECT raw_events_first.user_id 
- FROM   raw_events_first, 
-        raw_events_second 
- WHERE  raw_events_second.user_id = raw_events_first.value_1 
+ INSERT INTO agg_events
+             (user_id)
+ SELECT raw_events_first.user_id
+ FROM   raw_events_first,
+        raw_events_second
+ WHERE  raw_events_second.user_id = raw_events_first.value_1
         AND raw_events_first.value_2 = 12;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
  -- lets do some unsupported query tests with subqueries
- -- foo is not joined on the partition key so the query is not 
+ -- foo is not joined on the partition key so the query is not
  -- pushed down
  INSERT INTO agg_events
              (user_id, value_4_agg)
@@ -1177,7 +1158,7 @@ INSERT INTO agg_events
 SELECT SUM(value_3),
        Count(value_4),
        user_id,
-       SUM(value_1), 
+       SUM(value_1),
        value_2
 FROM   raw_events_first
 GROUP  BY user_id,
@@ -1438,7 +1419,7 @@ ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 -- not supported subqueries in WHERE clause
 -- since the selected value in the WHERE is not
--- partition key  
+-- partition key
 INSERT INTO raw_events_second
             (user_id)
 SELECT user_id
@@ -1525,7 +1506,7 @@ DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300005 AS
 DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300006 AS citus_table_alias (user_id) SELECT user_id FROM public.raw_events_first_13300002 raw_events_first WHERE ((NOT (EXISTS (SELECT 1 FROM public.raw_events_second_13300006 raw_events_second WHERE (raw_events_second.user_id OPERATOR(pg_catalog.=) raw_events_first.user_id)))) AND ((worker_hash(user_id) OPERATOR(pg_catalog.>=) 0) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) 1073741823)))
 DEBUG:  distributed statement: INSERT INTO public.raw_events_second_13300007 AS citus_table_alias (user_id) SELECT user_id FROM public.raw_events_first_13300003 raw_events_first WHERE ((NOT (EXISTS (SELECT 1 FROM public.raw_events_second_13300007 raw_events_second WHERE (raw_events_second.user_id OPERATOR(pg_catalog.=) raw_events_first.user_id)))) AND ((worker_hash(user_id) OPERATOR(pg_catalog.>=) 1073741824) AND (worker_hash(user_id) OPERATOR(pg_catalog.<=) 2147483647)))
 DEBUG:  Plan is router executable
--- more complex LEFT JOINs 
+-- more complex LEFT JOINs
  INSERT INTO agg_events
              (user_id, value_4_agg)
  SELECT
@@ -1705,39 +1686,39 @@ TRUNCATE raw_events_first;
 -- we don't support LIMIT for subquery pushdown, but
 -- we recursively plan the query and run it via coordinator
 INSERT INTO agg_events(user_id)
-SELECT user_id 
-FROM   users_table 
-WHERE  user_id 
-  IN (SELECT 
-      user_id 
+SELECT user_id
+FROM   users_table
+WHERE  user_id
+  IN (SELECT
+      user_id
         FROM  (
                 (
-                  SELECT 
-                    user_id 
+                  SELECT
+                    user_id
                     FROM
                     (
-                      SELECT 
-                        e1.user_id 
-                      FROM 
-                        users_table u1, events_table e1 
-                      WHERE 
+                      SELECT
+                        e1.user_id
+                      FROM
+                        users_table u1, events_table e1
+                      WHERE
                         e1.user_id = u1.user_id LIMIT 3
                      ) as f_inner
                   )
-          ) AS f2); 
+          ) AS f2);
 -- Altering a table and selecting from it using a multi-shard statement
 -- in the same transaction is allowed because we will use the same
 -- connections for all co-located placements.
 BEGIN;
 ALTER TABLE raw_events_second DROP COLUMN value_4;
-INSERT INTO raw_events_first SELECT * FROM raw_events_second; 
+INSERT INTO raw_events_first SELECT * FROM raw_events_second;
 ROLLBACK;
 -- Alterating a table and selecting from it using a single-shard statement
 -- in the same transaction is disallowed because we will use a different
 -- connection.
 BEGIN;
 ALTER TABLE raw_events_second DROP COLUMN value_4;
-INSERT INTO raw_events_first SELECT * FROM raw_events_second WHERE user_id = 100; 
+INSERT INTO raw_events_first SELECT * FROM raw_events_second WHERE user_id = 100;
 ROLLBACK;
 -- Altering a reference table and then performing an INSERT ... SELECT which
 -- joins with the reference table is allowed, since the INSERT ... SELECT
@@ -1897,17 +1878,17 @@ INSERT INTO insert_select_varchar_test (key, value)
 SELECT *, 100
 FROM   (SELECT f1.key
         FROM   (SELECT key
-                FROM   insert_select_varchar_test 
+                FROM   insert_select_varchar_test
                 GROUP  BY 1
-                HAVING Count(key) < 3) AS f1, 
-               (SELECT key 
-                FROM   insert_select_varchar_test 
-                GROUP  BY 1 
-                HAVING Sum(COALESCE(insert_select_varchar_test.value, 0)) > 
-                       20.0) 
-               AS f2 
-        WHERE  f1.key = f2.key 
-        GROUP  BY 1) AS foo; 
+                HAVING Count(key) < 3) AS f1,
+               (SELECT key
+                FROM   insert_select_varchar_test
+                GROUP  BY 1
+                HAVING Sum(COALESCE(insert_select_varchar_test.value, 0)) >
+                       20.0)
+               AS f2
+        WHERE  f1.key = f2.key
+        GROUP  BY 1) AS foo;
 SELECT * FROM insert_select_varchar_test ORDER BY 1 DESC, 2 DESC;
   key   | value 
 --------+-------
@@ -1920,7 +1901,7 @@ SELECT * FROM insert_select_varchar_test ORDER BY 1 DESC, 2 DESC;
 -- this test is mostly importantly intended for deparsing the query correctly
 -- but still it is preferable to have this test here instead of multi_deparse_shard_query
 CREATE TABLE table_with_defaults
-( 
+(
   store_id int,
   first_name text,
   default_1 int DEFAULT 1,
@@ -2101,10 +2082,10 @@ SELECT create_distributed_table('table_with_starts_with_defaults', 'c');
 (1 row)
 
 SET client_min_messages TO DEBUG;
-INSERT INTO text_table (part_col) 
-  SELECT 
+INSERT INTO text_table (part_col)
+  SELECT
     CASE WHEN part_col = 'onder' THEN 'marco'
-      END 
+      END
 FROM text_table ;
 DEBUG:  cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
 DETAIL:  Subquery contains a case expression in the same position as the target table's partition column.
@@ -2166,7 +2147,7 @@ DEBUG:  Collecting INSERT ... SELECT results on coordinator
 DEBUG:  Router planner cannot handle multi-shard select queries
 RESET client_min_messages;
 insert into table_with_starts_with_defaults (b,c) select b,c FROM table_with_starts_with_defaults;
--- Test on partition column without native hash function 
+-- Test on partition column without native hash function
 CREATE TABLE raw_table
 (
     id BIGINT,
@@ -2774,9 +2755,9 @@ SELECT * FROM coerce_agg ORDER BY 1 DESC, 2 DESC;
 
 -- INSERT..SELECT + prepared statements + recursive planning
 BEGIN;
-PREPARE prepared_recursive_insert_select AS 
-INSERT INTO users_table 
-SELECT * FROM users_table 
+PREPARE prepared_recursive_insert_select AS
+INSERT INTO users_table
+SELECT * FROM users_table
 WHERE value_1 IN (SELECT value_2 FROM events_table OFFSET 0);
 EXECUTE prepared_recursive_insert_select;
 EXECUTE prepared_recursive_insert_select;

src/test/regress/expected/multi_insert_select_non_pushable_queries.out

@@ -4,7 +4,7 @@
 ------------------------------------
 ------------------------------------
 CREATE TABLE test_table_1(id int);
-INSERT INTO test_table_1 
+INSERT INTO test_table_1
 SELECT user_id FROM users_table;
 ERROR:  cannot INSERT rows from a distributed query into a local table
 HINT:  Consider using CREATE TEMPORARY TABLE tmp AS SELECT ... and inserting from the temporary table.
@@ -67,10 +67,9 @@ FROM (
       AND e.event_type IN (103, 104, 105)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
-      
       WHERE  e.user_id >= 10
       AND e.user_id <= 25
       AND e.event_type IN (106, 107, 108)
@@ -108,10 +107,9 @@ FROM (
       AND e.event_type IN (103, 104, 105)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
-      
       WHERE  e.user_id >= 10
       AND e.user_id <= 25
       AND e.event_type IN (106, 107, 108)
@@ -151,10 +149,9 @@ FROM (
       AND e.event_type IN (103, 104, 105)
     )
   ) t1 LEFT JOIN (
-      SELECT DISTINCT user_id, 
+      SELECT DISTINCT user_id,
         'Has done event'::TEXT AS hasdone_event
       FROM  events_table AS e
-      
       WHERE  e.user_id >= 10
       AND e.user_id <= 25
       AND e.event_type IN (106, 107, 108)
@@ -225,7 +222,7 @@ SELECT
       users_table
     WHERE
       user_id >= 10 AND
-      user_id <= 70 AND    
+      user_id <= 70 AND
       users_table.value_1 > 15 AND users_table.value_1 < 17
     GROUP BY
       user_id
@@ -295,7 +292,7 @@ SELECT
       users_table
     WHERE
       user_id >= 10 AND
-      user_id <= 70 AND    
+      user_id <= 70 AND
       users_table.value_1 > 15 AND users_table.value_1 < 17
     GROUP BY
       user_id
@@ -340,7 +337,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time
         FROM users_table
         WHERE
@@ -370,7 +366,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time
         FROM users_table
         WHERE
@@ -400,7 +395,6 @@ FROM (
         max(u.time) as user_lastseen,
         array_agg(event_type ORDER BY u.time) AS event_array
     FROM (
-        
         SELECT user_id, time, value_3 as val_3
         FROM users_table
         WHERE
@@ -519,7 +513,6 @@ SELECT user_id, value_2 FROM users_table WHERE
   AND  NOT EXISTS (SELECT user_id FROM events_table WHERE event_type > 300 AND event_type <= 350  AND value_3 > 100 AND user_id=users_table.user_id);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
-  
 ------------------------------------
 ------------------------------------
 -- Customers who have done X more than 2 times, and satisfy other customer specific criteria
@@ -527,55 +520,55 @@ DETAIL:  Select query cannot be pushed down to the worker.
 ------------------------------------
 -- not pushable since the second join is not an equi join
 INSERT INTO agg_results_third(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
-  WHERE  value_1 > 100 
-         AND value_1 < 124 
-         AND value_2 >= 5 
-         AND EXISTS (SELECT user_id 
+  WHERE  value_1 > 100
+         AND value_1 < 124
+         AND value_2 >= 5
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type > 100 
-                            AND event_type < 124 
-                            AND value_3 > 100 
-                            AND user_id != users_table.user_id 
-                     GROUP  BY user_id 
+                     WHERE  event_type > 100
+                            AND event_type < 124
+                            AND value_3 > 100
+                            AND user_id != users_table.user_id
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 -- not pushable since the second join is not on the partition key
 INSERT INTO agg_results_third(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
-  WHERE  value_1 > 100 
-         AND value_1 < 124 
-         AND value_2 >= 5 
-         AND EXISTS (SELECT user_id 
+  WHERE  value_1 > 100
+         AND value_1 < 124
+         AND value_2 >= 5
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type > 100 
-                            AND event_type < 124 
-                            AND value_3 > 100 
-                            AND event_type = users_table.user_id 
-                     GROUP  BY user_id 
+                     WHERE  event_type > 100
+                            AND event_type < 124
+                            AND value_3 > 100
+                            AND event_type = users_table.user_id
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 -- not pushable since the second join is not on the partition key
 INSERT INTO agg_results_third(user_id, value_2_agg)
-  SELECT user_id, 
-         value_2 
+  SELECT user_id,
+         value_2
   FROM   users_table
-  WHERE  value_1 > 100 
-         AND value_1 < 124 
-         AND value_2 >= 5 
-         AND EXISTS (SELECT user_id 
+  WHERE  value_1 > 100
+         AND value_1 < 124
+         AND value_2 >= 5
+         AND EXISTS (SELECT user_id
                      FROM   events_table
-                     WHERE  event_type > 100 
-                            AND event_type < 124 
-                            AND value_3 > 100 
-                            AND user_id = users_table.value_1 
-                     GROUP  BY user_id 
+                     WHERE  event_type > 100
+                            AND event_type < 124
+                            AND value_3 > 100
+                            AND user_id = users_table.value_1
+                     GROUP  BY user_id
                      HAVING Count(*) > 2);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
@@ -594,7 +587,7 @@ And user_id NOT in
   (select user_id
    From users_table
    Where value_1 = 15
-   And value_2 > 25); 
+   And value_2 > 25);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 -- not pushable since we're not selecting the partition key
@@ -607,10 +600,9 @@ And user_id  in
   (select value_3
    From users_table
    Where value_1 = 15
-   And value_2 > 25);  
+   And value_2 > 25);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
- 
  -- not pushable since we're not selecting the partition key
  -- from the events table
 INSERT INTO agg_results_third(user_id)
@@ -622,7 +614,7 @@ And event_type  in
   (select user_id
    From users_table
    Where value_1 = 15
-   And value_2 > 25);  
+   And value_2 > 25);
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 ------------------------------------
@@ -660,17 +652,17 @@ DETAIL:  Select query cannot be pushed down to the worker.
 INSERT INTO agg_results_third(user_id, value_1_agg, value_3_agg)
 SELECT
     users_table.user_id, users_table.value_1, prob
-FROM 
+FROM
    users_table
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, (GREATEST(coalesce(ma.value_4 / 250, 0.0) + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
       users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
       short_list.user_id != ma.user_id and ma.value_1 < 50 and short_list.event_type < 50
-    ) temp 
-  ON users_table.user_id = temp.user_id 
+    ) temp
+  ON users_table.user_id = temp.user_id
   WHERE users_table.value_1 < 50;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
@@ -678,17 +670,17 @@ DETAIL:  Select query cannot be pushed down to the worker.
 INSERT INTO agg_results_third(user_id, value_1_agg, value_3_agg)
 SELECT
     users_table.user_id, users_table.value_1, prob
-FROM 
+FROM
    users_table
-        JOIN 
-   (SELECT  
+        JOIN
+   (SELECT
       ma.user_id, (GREATEST(coalesce(ma.value_4 / 250, 0.0) + GREATEST(1.0))) / 2 AS prob
-    FROM 
+    FROM
       users_table AS ma, events_table as short_list
-    WHERE 
+    WHERE
       short_list.user_id = ma.value_2 and ma.value_1 < 50 and short_list.event_type < 50
-    ) temp 
-  ON users_table.user_id = temp.user_id 
+    ) temp
+  ON users_table.user_id = temp.user_id
   WHERE users_table.value_1 < 50;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.

src/test/regress/expected/multi_json_agg.out

@@ -10,7 +10,7 @@ SELECT json_agg(value) FROM (
 ) t
 $$;
 -- Check multi_cat_agg() aggregate which is used to implement json_agg()
-SELECT json_cat_agg(i) FROM 
+SELECT json_cat_agg(i) FROM
 	(VALUES ('[1,{"a":2}]'::json), ('[null]'::json), (NULL), ('["3",5,4]'::json)) AS t(i);
          json_cat_agg          
 -------------------------------
@@ -141,7 +141,6 @@ SELECT json_agg(case when l_quantity > 20 then l_quantity else NULL end)
  [null, 36.00, null, 28.00, 24.00, 32.00, 38.00, 45.00, 49.00, 27.00, null, 28.00, 26.00, 30.00]
 (1 row)
 
-	
 -- Check that we can execute json_agg() with an expression containing different types
 SELECT json_agg(case when l_quantity > 20 then to_json(l_quantity) else '"f"'::json end)
 	FROM lineitem WHERE l_orderkey < 5;
@@ -150,7 +149,6 @@ SELECT json_agg(case when l_quantity > 20 then to_json(l_quantity) else '"f"'::j
  ["f", 36.00, "f", 28.00, 24.00, 32.00, 38.00, 45.00, 49.00, 27.00, "f", 28.00, 26.00, 30.00]
 (1 row)
 
-	
 -- Check that we can execute json_agg() with an expression containing json arrays
 SELECT json_agg(json_build_array(l_quantity, l_shipdate))
 	FROM lineitem WHERE l_orderkey < 3;
@@ -159,7 +157,6 @@ SELECT json_agg(json_build_array(l_quantity, l_shipdate))
  [[17.00, "1996-03-13"], [36.00, "1996-04-12"], [8.00, "1996-01-29"], [28.00, "1996-04-21"], [24.00, "1996-03-30"], [32.00, "1996-01-30"], [38.00, "1997-01-28"]]
 (1 row)
 
-	
 -- Check that we can execute json_agg() with an expression containing arrays
 SELECT json_agg(ARRAY[l_quantity, l_orderkey])
 	FROM lineitem WHERE l_orderkey < 3;

src/test/regress/expected/multi_json_object_agg.out

@@ -15,7 +15,7 @@ SELECT json_object_agg(key, value) FROM (
 ) t
 $$;
 -- Check multi_cat_agg() aggregate which is used to implement json_object_agg()
-SELECT json_cat_agg(i) FROM 
+SELECT json_cat_agg(i) FROM
 	(VALUES ('{"c":[], "b":2}'::json), (NULL), ('{"d":null, "a":{"b":3}, "b":2}'::json)) AS t(i);
                        json_cat_agg                        
 -----------------------------------------------------------
@@ -30,7 +30,7 @@ ERROR:  json_object_agg with order by is unsupported
 SELECT json_object_agg(distinct l_orderkey, l_shipmode ORDER BY l_orderkey) FROM lineitem;
 ERROR:  json_object_agg with order by is unsupported
 -- Check json_object_agg() for different data types and LIMIT clauses
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_partkey)) 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_partkey))
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                  keys_sort                                                 
@@ -47,7 +47,7 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_partk
  { "341" : 88362, "342" : 89414, "343" : 169544 }
 (10 rows)
 
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_extendedprice)) 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_extendedprice))
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                         keys_sort                                                         
@@ -64,7 +64,7 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_exten
  { "341" : 17554.68, "342" : 30875.02, "343" : 9681.24 }
 (10 rows)
 
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipmode)) 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipmode))
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                                        keys_sort                                                                       
@@ -81,7 +81,7 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipm
  { "341" : "REG AIR   ", "342" : "FOB       ", "343" : "FOB       " }
 (10 rows)
 
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate)) 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate))
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                                        keys_sort                                                                       
@@ -109,8 +109,8 @@ SELECT count_keys(json_object_agg(l_shipdate, l_orderkey)) FROM lineitem;
 -- shards and contain different aggregates, filter clauses and other complex
 -- expressions. Note that the l_orderkey ranges are such that the matching rows
 -- lie in different shards.
-SELECT l_quantity, count(*), avg(l_extendedprice), 
-		keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate)) 
+SELECT l_quantity, count(*), avg(l_extendedprice),
+		keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate))
 	FROM lineitem
 	WHERE l_quantity < 5 AND l_orderkey > 5000 AND l_orderkey < 5300
 	GROUP BY l_quantity ORDER BY l_quantity;
@@ -122,7 +122,7 @@ SELECT l_quantity, count(*), avg(l_extendedprice),
        4.00 |     4 | 5795.6800000000000000 | { "50313" : "1994-12-26", "50622" : "1993-02-06", "50891" : "1992-09-18", "51893" : "1993-12-21" }
 (4 rows)
 
-SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text,
 											extract (month FROM o_orderdate)))
 	FROM lineitem, orders WHERE l_orderkey = o_orderkey AND l_quantity < 5
 	AND l_orderkey > 5000 AND l_orderkey < 5300 GROUP BY l_quantity ORDER BY l_quantity;
@@ -134,7 +134,7 @@ SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::t
        4.00 | { "50313" : 12, "50622" : 10, "50891" : 7, "51893" : 11 }
 (4 rows)
 
-SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_orderkey * 2 + 1)) 
+SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, l_orderkey * 2 + 1))
 	FROM lineitem WHERE l_quantity < 5
 	AND octet_length(l_comment) + octet_length('randomtext'::text) > 40
 	AND l_orderkey > 5000 AND l_orderkey < 6000 GROUP BY l_quantity ORDER BY l_quantity;
@@ -147,7 +147,7 @@ SELECT l_quantity, keys_sort(json_object_agg(l_orderkey::text || l_linenumber::t
 (4 rows)
 
 -- Check that we can execute json_object_agg() with an expression containing NULL values
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text,
 								case when l_quantity > 20 then l_quantity else NULL end))
 	FROM lineitem WHERE l_orderkey < 5;
                                                                                               keys_sort                                                                                              
@@ -155,9 +155,8 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text,
  { "11" : null, "12" : 36.00, "13" : null, "14" : 28.00, "15" : 24.00, "16" : 32.00, "21" : 38.00, "31" : 45.00, "32" : 49.00, "33" : 27.00, "34" : null, "35" : 28.00, "36" : 26.00, "41" : 30.00 }
 (1 row)
 
-	
 -- Check that we can execute json_object_agg() with an expression containing different types
-SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text,
 								case when l_quantity > 20 then to_json(l_quantity) else '"f"'::json end))
 	FROM lineitem WHERE l_orderkey < 5;
                                                                                             keys_sort                                                                                             
@@ -165,7 +164,6 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text,
  { "11" : "f", "12" : 36.00, "13" : "f", "14" : 28.00, "15" : 24.00, "16" : 32.00, "21" : 38.00, "31" : 45.00, "32" : 49.00, "33" : 27.00, "34" : "f", "35" : 28.00, "36" : 26.00, "41" : 30.00 }
 (1 row)
 
-	
 -- Check that we can execute json_object_agg() with an expression containing json arrays
 SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, json_build_array(l_quantity, l_shipdate)))
 	FROM lineitem WHERE l_orderkey < 3;
@@ -174,7 +172,6 @@ SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, json_bu
  { "11" : [17.00, "1996-03-13"], "12" : [36.00, "1996-04-12"], "13" : [8.00, "1996-01-29"], "14" : [28.00, "1996-04-21"], "15" : [24.00, "1996-03-30"], "16" : [32.00, "1996-01-30"], "21" : [38.00, "1997-01-28"] }
 (1 row)
 
-	
 -- Check that we can execute json_object_agg() with an expression containing arrays
 SELECT keys_sort(json_object_agg(l_orderkey::text || l_linenumber::text, ARRAY[l_quantity, l_orderkey]))
 	FROM lineitem WHERE l_orderkey < 3;

src/test/regress/expected/multi_jsonb_agg.out

@@ -10,7 +10,7 @@ SELECT jsonb_agg(value) FROM (
 ) t
 $$;
 -- Check multi_cat_agg() aggregate which is used to implement jsonb_agg()
-SELECT jsonb_cat_agg(i) FROM 
+SELECT jsonb_cat_agg(i) FROM
 	(VALUES ('[1,{"a":2}]'::jsonb), ('[null]'::jsonb), (NULL), ('["3",5,4]'::jsonb)) AS t(i);
          jsonb_cat_agg          
 --------------------------------
@@ -141,7 +141,6 @@ SELECT jsonb_agg(case when l_quantity > 20 then l_quantity else NULL end)
  [null, 36.00, null, 28.00, 24.00, 32.00, 38.00, 45.00, 49.00, 27.00, null, 28.00, 26.00, 30.00]
 (1 row)
 
-	
 -- Check that we can execute jsonb_agg() with an expression containing different types
 SELECT jsonb_agg(case when l_quantity > 20 then to_jsonb(l_quantity) else '"f"'::jsonb end)
 	FROM lineitem WHERE l_orderkey < 5;
@@ -150,7 +149,6 @@ SELECT jsonb_agg(case when l_quantity > 20 then to_jsonb(l_quantity) else '"f"':
  ["f", 36.00, "f", 28.00, 24.00, 32.00, 38.00, 45.00, 49.00, 27.00, "f", 28.00, 26.00, 30.00]
 (1 row)
 
-	
 -- Check that we can execute jsonb_agg() with an expression containing jsonb arrays
 SELECT jsonb_agg(jsonb_build_array(l_quantity, l_shipdate))
 	FROM lineitem WHERE l_orderkey < 3;
@@ -159,7 +157,6 @@ SELECT jsonb_agg(jsonb_build_array(l_quantity, l_shipdate))
  [[17.00, "1996-03-13"], [36.00, "1996-04-12"], [8.00, "1996-01-29"], [28.00, "1996-04-21"], [24.00, "1996-03-30"], [32.00, "1996-01-30"], [38.00, "1997-01-28"]]
 (1 row)
 
-	
 -- Check that we can execute jsonb_agg() with an expression containing arrays
 SELECT jsonb_agg(ARRAY[l_quantity, l_orderkey])
 	FROM lineitem WHERE l_orderkey < 3;

src/test/regress/expected/multi_jsonb_object_agg.out

@@ -8,7 +8,7 @@ AS $$
 SELECT count(*) FROM (SELECT * FROM jsonb_object_keys($1)) t
 $$;
 -- Check multi_cat_agg() aggregate which is used to implement jsonb_object_agg()
-SELECT jsonb_cat_agg(i) FROM 
+SELECT jsonb_cat_agg(i) FROM
 	(VALUES ('{"c":[], "b":2}'::jsonb), (NULL), ('{"d":null, "a":{"b":3}, "b":2}'::jsonb)) AS t(i);
                 jsonb_cat_agg                
 ---------------------------------------------
@@ -23,7 +23,7 @@ ERROR:  jsonb_object_agg with order by is unsupported
 SELECT jsonb_object_agg(distinct l_orderkey, l_shipmode ORDER BY l_orderkey) FROM lineitem;
 ERROR:  jsonb_object_agg with order by is unsupported
 -- Check jsonb_object_agg() for different data types and LIMIT clauses
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_partkey) 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_partkey)
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                          jsonb_object_agg                                         
@@ -40,7 +40,7 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_partkey)
  {"341": 88362, "342": 89414, "343": 169544}
 (10 rows)
 
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_extendedprice) 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_extendedprice)
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                 jsonb_object_agg                                                 
@@ -57,7 +57,7 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_extendedprice)
  {"341": 17554.68, "342": 30875.02, "343": 9681.24}
 (10 rows)
 
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipmode) 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipmode)
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                                jsonb_object_agg                                                               
@@ -74,7 +74,7 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipmode)
  {"341": "REG AIR   ", "342": "FOB       ", "343": "FOB       "}
 (10 rows)
 
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate) 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate)
 	FROM lineitem GROUP BY l_orderkey
 	ORDER BY l_orderkey LIMIT 10;
                                                                jsonb_object_agg                                                               
@@ -102,8 +102,8 @@ SELECT count_keys(jsonb_object_agg(l_shipdate, l_orderkey)) FROM lineitem;
 -- shards and contain different aggregates, filter clauses and other complex
 -- expressions. Note that the l_orderkey ranges are such that the matching rows
 -- lie in different shards.
-SELECT l_quantity, count(*), avg(l_extendedprice), 
-		jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate) 
+SELECT l_quantity, count(*), avg(l_extendedprice),
+		jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_shipdate)
 	FROM lineitem
 	WHERE l_quantity < 5 AND l_orderkey > 5000 AND l_orderkey < 5300
 	GROUP BY l_quantity ORDER BY l_quantity;
@@ -115,7 +115,7 @@ SELECT l_quantity, count(*), avg(l_extendedprice),
        4.00 |     4 | 5795.6800000000000000 | {"50313": "1994-12-26", "50622": "1993-02-06", "50891": "1992-09-18", "51893": "1993-12-21"}
 (4 rows)
 
-SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text,
 									extract (month FROM o_orderdate))
 	FROM lineitem, orders WHERE l_orderkey = o_orderkey AND l_quantity < 5
 	AND l_orderkey > 5000 AND l_orderkey < 5300 GROUP BY l_quantity ORDER BY l_quantity;
@@ -127,7 +127,7 @@ SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text,
        4.00 | {"50313": 12, "50622": 10, "50891": 7, "51893": 11}
 (4 rows)
 
-SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_orderkey * 2 + 1) 
+SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_orderkey * 2 + 1)
 	FROM lineitem WHERE l_quantity < 5
 	AND octet_length(l_comment) + octet_length('randomtext'::text) > 40
 	AND l_orderkey > 5000 AND l_orderkey < 6000 GROUP BY l_quantity ORDER BY l_quantity;
@@ -140,7 +140,7 @@ SELECT l_quantity, jsonb_object_agg(l_orderkey::text || l_linenumber::text, l_or
 (4 rows)
 
 -- Check that we can execute jsonb_object_agg() with an expression containing NULL values
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text,
 						case when l_quantity > 20 then l_quantity else NULL end)
 	FROM lineitem WHERE l_orderkey < 5;
                                                                                   jsonb_object_agg                                                                                   
@@ -148,9 +148,8 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text,
  {"11": null, "12": 36.00, "13": null, "14": 28.00, "15": 24.00, "16": 32.00, "21": 38.00, "31": 45.00, "32": 49.00, "33": 27.00, "34": null, "35": 28.00, "36": 26.00, "41": 30.00}
 (1 row)
 
-	
 -- Check that we can execute jsonb_object_agg() with an expression containing different types
-SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, 
+SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text,
 						case when l_quantity > 20 then to_jsonb(l_quantity) else '"f"'::jsonb end)
 	FROM lineitem WHERE l_orderkey < 5;
                                                                                  jsonb_object_agg                                                                                 
@@ -158,7 +157,6 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text,
  {"11": "f", "12": 36.00, "13": "f", "14": 28.00, "15": 24.00, "16": 32.00, "21": 38.00, "31": 45.00, "32": 49.00, "33": 27.00, "34": "f", "35": 28.00, "36": 26.00, "41": 30.00}
 (1 row)
 
-	
 -- Check that we can execute jsonb_object_agg() with an expression containing jsonb arrays
 SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, jsonb_build_array(l_quantity, l_shipdate))
 	FROM lineitem WHERE l_orderkey < 3;
@@ -167,7 +165,6 @@ SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, jsonb_build_arra
  {"11": [17.00, "1996-03-13"], "12": [36.00, "1996-04-12"], "13": [8.00, "1996-01-29"], "14": [28.00, "1996-04-21"], "15": [24.00, "1996-03-30"], "16": [32.00, "1996-01-30"], "21": [38.00, "1997-01-28"]}
 (1 row)
 
-	
 -- Check that we can execute jsonb_object_agg() with an expression containing arrays
 SELECT jsonb_object_agg(l_orderkey::text || l_linenumber::text, ARRAY[l_quantity, l_orderkey])
 	FROM lineitem WHERE l_orderkey < 3;

src/test/regress/expected/multi_mx_partitioning.out

@@ -15,7 +15,6 @@ SELECT start_metadata_sync_to_node('localhost', :worker_1_port);
 -- 1-) Distributing partitioned table
 -- create partitioned table
 CREATE TABLE partitioning_test(id int, time date) PARTITION BY RANGE (time);
- 
 -- create its partitions
 CREATE TABLE partitioning_test_2009 PARTITION OF partitioning_test FOR VALUES FROM ('2009-01-01') TO ('2010-01-01');
 CREATE TABLE partitioning_test_2010 PARTITION OF partitioning_test FOR VALUES FROM ('2010-01-01') TO ('2011-01-01');
@@ -45,11 +44,11 @@ SELECT * FROM partitioning_test ORDER BY 1;
 (4 rows)
 
 -- see from MX node, partitioned table and its partitions are distributed
-SELECT 
-	logicalrelid 
-FROM 
-	pg_dist_partition 
-WHERE 
+SELECT
+	logicalrelid
+FROM
+	pg_dist_partition
+WHERE
 	logicalrelid IN ('partitioning_test', 'partitioning_test_2009', 'partitioning_test_2010')
 ORDER BY 1;
       logicalrelid      
@@ -59,9 +58,9 @@ ORDER BY 1;
  partitioning_test_2010
 (3 rows)
 
-SELECT 
-	logicalrelid, count(*) 
-FROM pg_dist_shard 
+SELECT
+	logicalrelid, count(*)
+FROM pg_dist_shard
 	WHERE logicalrelid IN ('partitioning_test', 'partitioning_test_2009', 'partitioning_test_2010')
 GROUP BY
 	logicalrelid
@@ -89,11 +88,11 @@ SET citus.shard_replication_factor TO 1;
 CREATE TABLE partitioning_test_2011 PARTITION OF partitioning_test FOR VALUES FROM ('2011-01-01') TO ('2012-01-01');
 -- see from MX node, new partition is automatically distributed as well
 \c - - - :worker_1_port
-SELECT 
-	logicalrelid 
-FROM 
-	pg_dist_partition 
-WHERE 
+SELECT
+	logicalrelid
+FROM
+	pg_dist_partition
+WHERE
 	logicalrelid IN ('partitioning_test', 'partitioning_test_2011')
 ORDER BY 1;
       logicalrelid      
@@ -102,9 +101,9 @@ ORDER BY 1;
  partitioning_test_2011
 (2 rows)
 
-SELECT 
-	logicalrelid, count(*) 
-FROM pg_dist_shard 
+SELECT
+	logicalrelid, count(*)
+FROM pg_dist_shard
 	WHERE logicalrelid IN ('partitioning_test', 'partitioning_test_2011')
 GROUP BY
 	logicalrelid
@@ -137,11 +136,11 @@ ALTER TABLE partitioning_test ATTACH PARTITION partitioning_test_2012 FOR VALUES
 NOTICE:  Copying data from local table...
 -- see from MX node, attached partition is distributed as well
 \c - - - :worker_1_port
-SELECT 
-	logicalrelid 
-FROM 
-	pg_dist_partition 
-WHERE 
+SELECT
+	logicalrelid
+FROM
+	pg_dist_partition
+WHERE
 	logicalrelid IN ('partitioning_test', 'partitioning_test_2012')
 ORDER BY 1;
       logicalrelid      
@@ -150,9 +149,9 @@ ORDER BY 1;
  partitioning_test_2012
 (2 rows)
 
-SELECT 
-	logicalrelid, count(*) 
-FROM pg_dist_shard 
+SELECT
+	logicalrelid, count(*)
+FROM pg_dist_shard
 	WHERE logicalrelid IN ('partitioning_test', 'partitioning_test_2012')
 GROUP BY
 	logicalrelid

src/test/regress/expected/multi_mx_repartition_udt_prepare.out

@@ -47,7 +47,7 @@ CREATE TABLE repartition_udt_other (
     udtcol test_udt,
     txtcol text
 );
--- Connect directly to a worker, create and drop the type, then 
+-- Connect directly to a worker, create and drop the type, then
 -- proceed with type creation as above; thus the OIDs will be different.
 -- so that the OID is off.
 \c - - - :worker_1_port
@@ -194,6 +194,5 @@ LOG:  join order: [ "repartition_udt" ][ dual partition join "repartition_udt_ot
   6 | (2,3)  | foo    | 12 | (2,3)  | foo
 (5 rows)
 
-    
 \c - - - :worker_1_port
 \c - - - :worker_2_port

src/test/regress/expected/multi_mx_router_planner.out

@@ -65,8 +65,8 @@ DEBUG:  Plan is router executable
 DETAIL:  distribution column value: 10
 -- single-shard tests
 -- many of the tests in this file is intended for testing non-fast-path
--- router planner, so we're explicitly disabling it in this file. 
--- We've bunch of other tests that triggers fast-path-router 
+-- router planner, so we're explicitly disabling it in this file.
+-- We've bunch of other tests that triggers fast-path-router
 SET citus.enable_fast_path_router_planner TO false;
 -- test simple select for a single row
 SELECT * FROM articles_hash_mx WHERE author_id = 10 AND id = 50;
@@ -195,7 +195,7 @@ DETAIL:  distribution column value: 1
 
 -- query is a single shard query but can't do shard pruning,
 -- not router-plannable due to <= and IN
-SELECT * FROM articles_hash_mx WHERE author_id <= 1; 
+SELECT * FROM articles_hash_mx WHERE author_id <= 1;
 DEBUG:  Router planner cannot handle multi-shard select queries
  id | author_id |    title     | word_count 
 ----+-----------+--------------+------------
@@ -206,7 +206,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-SELECT * FROM articles_hash_mx WHERE author_id IN (1, 3); 
+SELECT * FROM articles_hash_mx WHERE author_id IN (1, 3);
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
  id | author_id |    title     | word_count 
@@ -335,7 +335,7 @@ DETAIL:  distribution column value: 3
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees_mx
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees_mx ce
@@ -358,7 +358,7 @@ DETAIL:  distribution column value: 1
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees_mx
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees_mx ce
@@ -373,7 +373,7 @@ ERROR:  recursive CTEs are not supported in distributed queries
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees_mx
-		WHERE company_id = 3 and manager_id = 0 
+		WHERE company_id = 3 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees_mx ce
@@ -488,7 +488,7 @@ DEBUG:  push down of limit count: 5
 (5 rows)
 
 SELECT articles_hash_mx.id,test.word_count
-FROM articles_hash_mx, (SELECT id, word_count FROM articles_hash_mx) AS test 
+FROM articles_hash_mx, (SELECT id, word_count FROM articles_hash_mx) AS test
 WHERE test.id = articles_hash_mx.id and articles_hash_mx.author_id = 1
 ORDER BY articles_hash_mx.id;
 DEBUG:  Router planner cannot handle multi-shard select queries
@@ -591,7 +591,7 @@ DETAIL:  distribution column value: 10
            10 |              6363
 (3 rows)
 
--- following join is router plannable since the same worker 
+-- following join is router plannable since the same worker
 -- has both shards
 SELECT a.author_id as first_author, b.word_count as second_word_count
 	FROM articles_hash_mx a, articles_single_shard_hash_mx b
@@ -607,7 +607,6 @@ DETAIL:  distribution column value: 10
            10 |             19519
 (3 rows)
 
-	
 -- following join is not router plannable since there are no
 -- workers containing both shards, but will work through recursive
 -- planning
@@ -675,7 +674,6 @@ DETAIL:  distribution column value: 1
  21 |         1 | arcading     |       5890
 (2 rows)
 
-	
 -- single shard select with group by on non-partition column is router plannable
 SELECT id
 	FROM articles_hash_mx
@@ -749,7 +747,6 @@ DETAIL:  distribution column value: 1
  11814
 (1 row)
 
-	
 -- router plannable union queries are supported
 SELECT * FROM (
 	SELECT * FROM articles_hash_mx WHERE author_id = 1
@@ -913,7 +910,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- router plannable
 SELECT *
 	FROM articles_hash_mx
@@ -966,7 +962,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- router plannable due to abs(-1) getting converted to 1 by postgresql
 SELECT *
 	FROM articles_hash_mx
@@ -1123,7 +1118,7 @@ DETAIL:  distribution column value: 1
 (3 rows)
 
 -- window functions are supported if query is router plannable
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash_mx
 	WHERE author_id = 5;
 DEBUG:  Creating router plan
@@ -1138,7 +1133,7 @@ DETAIL:  distribution column value: 5
  aminate  | aruru    |      11389
 (5 rows)
 
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash_mx
 	WHERE author_id = 5
 	ORDER BY word_count DESC;
@@ -1184,8 +1179,8 @@ DETAIL:  distribution column value: 1
  41 |      11814 | 7178.8000000000000000
 (5 rows)
 
-SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)  
-	FROM articles_hash_mx 
+SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)
+	FROM articles_hash_mx
 	WHERE author_id = 1;
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
@@ -1206,16 +1201,16 @@ SELECT id, MIN(id) over (order by word_count)
 DEBUG:  Router planner cannot handle multi-shard select queries
 ERROR:  could not run distributed query because the window function that is used cannot be pushed down
 HINT:  Window functions are supported in two ways. Either add an equality filter on the distributed tables' partition column or use the window functions with a PARTITION BY clause containing the distribution column
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash_mx
 	WHERE author_id = 5 or author_id = 2;
 DEBUG:  Router planner cannot handle multi-shard select queries
 ERROR:  could not run distributed query because the window function that is used cannot be pushed down
 HINT:  Window functions are supported in two ways. Either add an equality filter on the distributed tables' partition column or use the window functions with a PARTITION BY clause containing the distribution column
--- complex query hitting a single shard 	
+-- complex query hitting a single shard
 SELECT
 	count(DISTINCT CASE
-			WHEN 
+			WHEN
 				word_count > 100
 			THEN
 				id
@@ -1237,7 +1232,7 @@ DETAIL:  distribution column value: 5
 -- same query is not router plannable if hits multiple shards
 SELECT
 	count(DISTINCT CASE
-			WHEN 
+			WHEN
 				word_count > 100
 			THEN
 				id
@@ -1285,7 +1280,7 @@ DETAIL:  distribution column value: 1
 END;
 -- cursor queries are router plannable
 BEGIN;
-DECLARE test_cursor CURSOR FOR 
+DECLARE test_cursor CURSOR FOR
 	SELECT *
 		FROM articles_hash_mx
 		WHERE author_id = 1
@@ -1326,7 +1321,6 @@ DETAIL:  distribution column value: 1
 21	1	arcading	5890
 31	1	athwartships	7271
 41	1	aznavour	11814
-	
 -- table creation queries inside can be router plannable
 CREATE TEMP TABLE temp_articles_hash_mx as
 	SELECT *
@@ -1474,7 +1468,6 @@ SET client_min_messages to 'DEBUG2';
 CREATE MATERIALIZED VIEW mv_articles_hash_mx_error AS
 	SELECT * FROM articles_hash_mx WHERE author_id in (1,2);
 DEBUG:  Router planner cannot handle multi-shard select queries
-	
 -- router planner/executor is disabled for task-tracker executor
 -- following query is router plannable, but router planner is disabled
 -- TODO: Uncomment once we fix task-tracker issue

src/test/regress/expected/multi_remove_node_reference_table.out

@@ -145,7 +145,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -164,7 +163,6 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
 \c - - - :master_port
 SELECT master_remove_node('localhost', :worker_2_port);
  master_remove_node 
@@ -217,7 +215,6 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
 \c - - - :master_port
 -- remove same node twice
 SELECT master_remove_node('localhost', :worker_2_port);
@@ -281,7 +278,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -300,7 +296,6 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
 \c - - - :master_port
 BEGIN;
 SELECT master_remove_node('localhost', :worker_2_port);
@@ -357,7 +352,6 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
 \c - - - :master_port
 -- remove node in a transaction and COMMIT
 -- status before master_remove_node
@@ -389,7 +383,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -408,7 +401,6 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
 \c - - - :master_port
 BEGIN;
 SELECT master_remove_node('localhost', :worker_2_port);
@@ -446,7 +438,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -464,7 +455,6 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
 \c - - - :master_port
 -- re-add the node for next tests
 SELECT 1 FROM master_add_node('localhost', :worker_2_port);
@@ -522,9 +512,7 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
-\c - - - :master_port     
-     
+\c - - - :master_port
 BEGIN;
 INSERT INTO remove_node_reference_table VALUES(1);
 SELECT master_remove_node('localhost', :worker_2_port);
@@ -586,14 +574,12 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
 SELECT * FROM remove_node_reference_table;
  column1 
 ---------
        1
 (1 row)
 
-    
 \c - - - :master_port
 -- re-add the node for next tests
 SELECT 1 FROM master_add_node('localhost', :worker_2_port);
@@ -651,7 +637,6 @@ WHERE
  1380000 |          1 |           0 | localhost |    57638
 (1 row)
 
-    
 \c - - - :master_port
 BEGIN;
 ALTER TABLE remove_node_reference_table ADD column2 int;
@@ -690,7 +675,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -708,7 +692,6 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
 \c - - - :master_port
 SET citus.next_shard_id TO 1380001;
 -- verify table structure is changed
@@ -844,7 +827,6 @@ WHERE colocationid IN
         10004 |          1 |                -1 |                      0
 (1 row)
 
-     
 \c - - - :worker_1_port
 SELECT COUNT(*) FROM pg_dist_node WHERE nodeport = :worker_2_port;
  count 
@@ -918,9 +900,7 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
-\c - - - :master_port     
-     
+\c - - - :master_port
 -- re-add the node for next tests
 SELECT 1 FROM master_add_node('localhost', :worker_2_port);
 NOTICE:  Replicating reference table "remove_node_reference_table" to the node localhost:57638
@@ -983,9 +963,7 @@ ORDER BY shardid ASC;
  1380002 |          1 |           0 | localhost |    57638
 (2 rows)
 
-    
-\c - - - :master_port     
-     
+\c - - - :master_port
 SELECT master_disable_node('localhost', :worker_2_port);
  master_disable_node 
 ---------------------
@@ -1037,7 +1015,6 @@ WHERE
 ---------+------------+-------------+----------+----------
 (0 rows)
 
-    
 \c - - - :master_port
 -- re-add the node for next tests
 SELECT 1 FROM master_activate_node('localhost', :worker_2_port);

src/test/regress/expected/multi_router_planner.out

@@ -255,7 +255,7 @@ DETAIL:  distribution column value: 1
 
 -- query is a single shard query but can't do shard pruning,
 -- not router-plannable due to <= and IN
-SELECT * FROM articles_hash WHERE author_id <= 1; 
+SELECT * FROM articles_hash WHERE author_id <= 1;
 DEBUG:  Router planner cannot handle multi-shard select queries
  id | author_id |    title     | word_count 
 ----+-----------+--------------+------------
@@ -266,7 +266,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-SELECT * FROM articles_hash WHERE author_id IN (1, 3); 
+SELECT * FROM articles_hash WHERE author_id IN (1, 3);
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
  id | author_id |    title     | word_count 
@@ -359,7 +359,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- recursive CTEs are supported when filtered on partition column
-CREATE TABLE company_employees (company_id int, employee_id int, manager_id int); 
+CREATE TABLE company_employees (company_id int, employee_id int, manager_id int);
 SELECT master_create_distributed_table('company_employees', 'company_id', 'hash');
  master_create_distributed_table 
 ---------------------------------
@@ -408,7 +408,7 @@ DETAIL:  distribution column value: 3
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -431,7 +431,7 @@ DETAIL:  distribution column value: 1
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -445,7 +445,7 @@ ERROR:  recursive CTEs are not supported in distributed queries
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 3 and manager_id = 0 
+		WHERE company_id = 3 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -647,7 +647,7 @@ DEBUG:  push down of limit count: 5
 (5 rows)
 
 SELECT articles_hash.id,test.word_count
-FROM articles_hash, (SELECT id, word_count FROM articles_hash) AS test 
+FROM articles_hash, (SELECT id, word_count FROM articles_hash) AS test
 WHERE test.id = articles_hash.id and articles_hash.author_id = 1
 ORDER BY articles_hash.id;
 DEBUG:  Router planner cannot handle multi-shard select queries
@@ -751,7 +751,7 @@ DETAIL:  distribution column value: 10
            10 |              6363
 (3 rows)
 
--- following join is router plannable since the same worker 
+-- following join is router plannable since the same worker
 -- has both shards
 SELECT a.author_id as first_author, b.word_count as second_word_count
 	FROM articles_hash a, articles_single_shard_hash b
@@ -767,7 +767,6 @@ DETAIL:  distribution column value: 10
            10 |             19519
 (3 rows)
 
-	
 -- following join is not router plannable since there are no
 -- workers containing both shards, but will work through recursive
 -- planning
@@ -835,7 +834,6 @@ DETAIL:  distribution column value: 1
  21 |         1 | arcading     |       5890
 (2 rows)
 
-	
 -- single shard select with group by on non-partition column is router plannable
 SELECT id
 	FROM articles_hash
@@ -1094,7 +1092,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- router plannable
 SELECT *
 	FROM articles_hash
@@ -1147,7 +1144,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- router plannable due to abs(-1) getting converted to 1 by postgresql
 SELECT *
 	FROM articles_hash
@@ -1304,7 +1300,7 @@ DETAIL:  distribution column value: 1
 (3 rows)
 
 -- window functions are supported if query is router plannable
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash
 	WHERE author_id = 5;
 DEBUG:  Creating router plan
@@ -1319,7 +1315,7 @@ DETAIL:  distribution column value: 5
  aminate  | aruru    |      11389
 (5 rows)
 
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash
 	WHERE author_id = 5
 	ORDER BY word_count DESC;
@@ -1365,8 +1361,8 @@ DETAIL:  distribution column value: 1
  41 |      11814 | 7178.8000000000000000
 (5 rows)
 
-SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)  
-	FROM articles_hash 
+SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)
+	FROM articles_hash
 	WHERE author_id = 1;
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
@@ -1387,14 +1383,14 @@ SELECT id, MIN(id) over (order by word_count)
 DEBUG:  Router planner cannot handle multi-shard select queries
 ERROR:  could not run distributed query because the window function that is used cannot be pushed down
 HINT:  Window functions are supported in two ways. Either add an equality filter on the distributed tables' partition column or use the window functions with a PARTITION BY clause containing the distribution column
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash
 	WHERE author_id = 5 or author_id = 2;
 DEBUG:  Router planner cannot handle multi-shard select queries
 ERROR:  could not run distributed query because the window function that is used cannot be pushed down
 HINT:  Window functions are supported in two ways. Either add an equality filter on the distributed tables' partition column or use the window functions with a PARTITION BY clause containing the distribution column
 -- where false queries are router plannable
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE false;
 DEBUG:  Creating router plan
@@ -1403,7 +1399,7 @@ DEBUG:  Plan is router executable
 ----+-----------+-------+------------
 (0 rows)
 
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE author_id = 1 and false;
 DEBUG:  Creating router plan
@@ -1412,7 +1408,7 @@ DEBUG:  Plan is router executable
 ----+-----------+-------+------------
 (0 rows)
 
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE author_id = 1 and 1=0;
 DEBUG:  Creating router plan
@@ -1430,7 +1426,7 @@ DEBUG:  Plan is router executable
 --------------+-------------------
 (0 rows)
 
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE null;
 DEBUG:  Creating router plan
@@ -1440,7 +1436,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- where false with immutable function returning false
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id = 10 and int4eq(1, 2);
 DEBUG:  Creating router plan
@@ -1449,7 +1445,7 @@ DEBUG:  Plan is router executable
 ----+-----------+-------+------------
 (0 rows)
 
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE int4eq(1, 2);
 DEBUG:  Creating router plan
@@ -1484,7 +1480,7 @@ DEBUG:  Plan is router executable
 
 -- partition_column is null clause does not prune out any shards,
 -- all shards remain after shard pruning, not router plannable
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id is null;
 DEBUG:  Router planner cannot handle multi-shard select queries
@@ -1494,7 +1490,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
 
 -- partition_column equals to null clause prunes out all shards
 -- no shards after shard pruning, router plannable
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id = null;
 DEBUG:  Creating router plan
@@ -1504,7 +1500,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- stable function returning bool
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE date_ne_timestamp('1954-04-11', '1954-04-11'::timestamp);
 DEBUG:  Creating router plan
@@ -1597,7 +1593,7 @@ DEBUG:  Plan is router executable
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -1615,7 +1611,7 @@ DETAIL:  distribution column value: 1
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -1650,8 +1646,8 @@ DETAIL:  distribution column value: 1
 (0 rows)
 
 -- window functions with where false
-SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)  
-	FROM articles_hash 
+SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)
+	FROM articles_hash
 	WHERE author_id = 1 and 1=0;
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
@@ -1661,7 +1657,7 @@ DEBUG:  Plan is router executable
 
 -- function calls in WHERE clause with non-relational arguments
 SELECT author_id FROM articles_hash
-	WHERE 
+	WHERE
 		substring('hello world', 1, 5) = 'hello'
 	ORDER BY
 		author_id
@@ -1675,7 +1671,7 @@ DEBUG:  push down of limit count: 1
 
 -- when expression evaluates to false
 SELECT author_id FROM articles_hash
-	WHERE 
+	WHERE
 		substring('hello world', 1, 4) = 'hello'
 	ORDER BY
 		author_id
@@ -1686,9 +1682,8 @@ DEBUG:  Plan is router executable
 -----------
 (0 rows)
 
-	
 -- verify range partitioned tables can be used in router plannable queries
--- just 4 shards to be created for each table to make sure 
+-- just 4 shards to be created for each table to make sure
 -- they are 'co-located' pairwise
 SET citus.shard_replication_factor TO 1;
 SELECT master_create_distributed_table('authors_range', 'id', 'range');
@@ -1743,7 +1738,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- single shard joins on range partitioned table are router plannable
-SELECT * FROM articles_range ar join authors_range au on (ar.author_id = au.id) 
+SELECT * FROM articles_range ar join authors_range au on (ar.author_id = au.id)
 	WHERE ar.author_id = 1;
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
@@ -1801,7 +1796,7 @@ DETAIL:  Creating dependency on merge taskId 20
 -- This query was intended to test "this is a bug, it is a single shard join
 -- query but not router plannable". To run it using repartition join logic we
 -- change the join columns.
-SELECT * FROM articles_range ar join authors_range au on (ar.title = au.name) 
+SELECT * FROM articles_range ar join authors_range au on (ar.title = au.name)
 	WHERE ar.author_id = 1 or au.id = 5;
 DEBUG:  Router planner cannot handle multi-shard select queries
 DEBUG:  join prunable for task partitionId 0 and 1
@@ -1838,7 +1833,7 @@ DETAIL:  Creating dependency on merge taskId 20
 
 RESET citus.task_executor_type;
 -- bogus query, join on non-partition column, but router plannable due to filters
-SELECT * FROM articles_range ar join authors_range au on (ar.id = au.id) 
+SELECT * FROM articles_range ar join authors_range au on (ar.id = au.id)
 	WHERE ar.author_id = 1 and au.id < 10;
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
@@ -1916,7 +1911,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
 ----+-----------+-------+------------+------+----
 (0 rows)
 
-	
 -- following is a bug, function should have been
 -- evaluated at master before going to worker
 -- need to use a append distributed table here
@@ -1948,7 +1942,7 @@ SET client_min_messages TO ERROR;
 \set VERBOSITY terse
 SELECT raise_failed_execution_router($$
 	SELECT author_id FROM articles_append
-		WHERE 
+		WHERE
 			substring('articles_append'::regclass::text, 1, 5) = 'hello'
 		ORDER BY
 			author_id
@@ -1958,7 +1952,7 @@ ERROR:  Task failed to execute
 -- same query with where false but evaluation left to worker
 SELECT raise_failed_execution_router($$
 	SELECT author_id FROM articles_append
-		WHERE 
+		WHERE
 			substring('articles_append'::regclass::text, 1, 4) = 'hello'
 		ORDER BY
 			author_id
@@ -1968,7 +1962,7 @@ ERROR:  Task failed to execute
 -- same query on router planner with where false but evaluation left to worker
 SELECT raise_failed_execution_router($$
 	SELECT author_id FROM articles_single_shard_hash
-		WHERE 
+		WHERE
 			substring('articles_single_shard_hash'::regclass::text, 1, 4) = 'hello'
 		ORDER BY
 			author_id
@@ -1977,7 +1971,7 @@ $$);
 ERROR:  Task failed to execute
 SELECT raise_failed_execution_router($$
 	SELECT author_id FROM articles_hash
-		WHERE 
+		WHERE
 			author_id = 1
 			AND substring('articles_hash'::regclass::text, 1, 5) = 'hello'
 		ORDER BY
@@ -1993,7 +1987,7 @@ BEGIN
     RETURN md5($1::text);
 END;
 $$ LANGUAGE 'plpgsql' IMMUTABLE;
--- not router plannable, returns all rows 
+-- not router plannable, returns all rows
 SELECT * FROM articles_hash
 	WHERE
 		someDummyFunction('articles_hash') = md5('articles_hash')
@@ -2024,10 +2018,10 @@ ERROR:  Task failed to execute
 SET client_min_messages TO 'NOTICE';
 DROP FUNCTION someDummyFunction(regclass);
 SET client_min_messages TO 'DEBUG2';
--- complex query hitting a single shard 	
+-- complex query hitting a single shard
 SELECT
 	count(DISTINCT CASE
-			WHEN 
+			WHEN
 				word_count > 100
 			THEN
 				id
@@ -2049,7 +2043,7 @@ DETAIL:  distribution column value: 5
 -- same query is not router plannable if hits multiple shards
 SELECT
 	count(DISTINCT CASE
-			WHEN 
+			WHEN
 				word_count > 100
 			THEN
 				id
@@ -2117,7 +2111,7 @@ DETAIL:  distribution column value: 1
 END;
 -- cursor queries are router plannable
 BEGIN;
-DECLARE test_cursor CURSOR FOR 
+DECLARE test_cursor CURSOR FOR
 	SELECT *
 		FROM articles_hash
 		WHERE author_id = 1
@@ -2166,7 +2160,6 @@ DETAIL:  distribution column value: 1
 21	1	arcading	5890
 31	1	athwartships	7271
 41	1	aznavour	11814
-	
 -- table creation queries inside can be router plannable
 CREATE TEMP TABLE temp_articles_hash as
 	SELECT *

src/test/regress/expected/multi_router_planner_fast_path.out

@@ -2,12 +2,12 @@ CREATE SCHEMA fast_path_router_select;
 SET search_path TO fast_path_router_select;
 SET citus.next_shard_id TO 1840000;
 -- all the tests in this file is intended for testing fast-path
--- router planner, so we're explicitly enabling itin this file. 
--- We've bunch of other tests that triggers non-fast-path-router 
+-- router planner, so we're explicitly enabling itin this file.
+-- We've bunch of other tests that triggers non-fast-path-router
 -- planner (note this is already true by default)
 SET citus.enable_fast_path_router_planner TO true;
 -- ===================================================================
--- test router planner functionality for via fast path on 
+-- test router planner functionality for via fast path on
 -- single shard select queries
 -- ===================================================================
 CREATE TABLE articles_hash (
@@ -161,7 +161,7 @@ DETAIL:  distribution column value: 1
 (1 row)
 
 -- fast path planner only support = operator
-SELECT * FROM articles_hash WHERE author_id <= 1; 
+SELECT * FROM articles_hash WHERE author_id <= 1;
 DEBUG:  Router planner cannot handle multi-shard select queries
  id | author_id |    title     | word_count 
 ----+-----------+--------------+------------
@@ -172,7 +172,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-SELECT * FROM articles_hash WHERE author_id IN (1, 3); 
+SELECT * FROM articles_hash WHERE author_id IN (1, 3);
 DEBUG:  Creating router plan
 DEBUG:  Plan is router executable
  id | author_id |    title     | word_count 
@@ -221,7 +221,7 @@ DETAIL:  distribution column value: 1
 (5 rows)
 
 -- this is a different case where each CTE is recursively planned and those goes
--- through the fast-path router planner, but the top level join is not  
+-- through the fast-path router planner, but the top level join is not
 WITH id_author AS ( SELECT id, author_id FROM articles_hash WHERE author_id = 1),
 id_title AS (SELECT id, title from articles_hash WHERE author_id = 2)
 SELECT * FROM id_author, id_title WHERE id_author.id = id_title.id;
@@ -243,7 +243,7 @@ DEBUG:  Plan is router executable
 ----+-----------+----+-------
 (0 rows)
 
-CREATE TABLE company_employees (company_id int, employee_id int, manager_id int); 
+CREATE TABLE company_employees (company_id int, employee_id int, manager_id int);
 SELECT master_create_distributed_table('company_employees', 'company_id', 'hash');
  master_create_distributed_table 
 ---------------------------------
@@ -296,7 +296,7 @@ DETAIL:  distribution column value: 3
 WITH RECURSIVE hierarchy as (
 	SELECT *, 1 AS level
 		FROM company_employees
-		WHERE company_id = 1 and manager_id = 0 
+		WHERE company_id = 1 and manager_id = 0
 	UNION
 	SELECT ce.*, (h.level+1)
 		FROM hierarchy h JOIN company_employees ce
@@ -428,7 +428,7 @@ DEBUG:  push down of limit count: 5
 (5 rows)
 
 SELECT articles_hash.id,test.word_count
-FROM articles_hash, (SELECT id, word_count FROM articles_hash) AS test 
+FROM articles_hash, (SELECT id, word_count FROM articles_hash) AS test
 WHERE test.id = articles_hash.id and articles_hash.author_id = 1
 ORDER BY articles_hash.id;
 DEBUG:  Router planner cannot handle multi-shard select queries
@@ -568,7 +568,6 @@ DETAIL:  distribution column value: 1
  21 |         1 | arcading     |       5890
 (2 rows)
 
-	
 -- single shard select with group by on non-partition column goes through fast-path planning
 SELECT id
 	FROM articles_hash
@@ -698,7 +697,7 @@ DETAIL:  distribution column value: 1
 
 -- Test various filtering options for router plannable check
 SET client_min_messages to 'DEBUG2';
--- cannot go through fast-path if there is 
+-- cannot go through fast-path if there is
 -- explicit coercion
 SELECT *
 	FROM articles_hash
@@ -715,7 +714,7 @@ DETAIL:  distribution column value: 1
  41 |         1 | aznavour     |      11814
 (5 rows)
 
--- can go through fast-path if there is 
+-- can go through fast-path if there is
 -- implicit coercion
 -- This doesn't work see the related issue
 -- reported https://github.com/citusdata/citus/issues/2605
@@ -765,7 +764,6 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- goes through fast-path planning because
 -- the dist. key is ANDed with the rest of the
 -- filters
@@ -822,9 +820,8 @@ DEBUG:  Router planner cannot handle multi-shard select queries
  41 |         1 | aznavour     |      11814
 (5 rows)
 
-	
 -- Citus does not qualify this as a fast-path because
--- dist_key = func() 
+-- dist_key = func()
 SELECT *
 	FROM articles_hash
 	WHERE author_id = abs(-1);
@@ -841,7 +838,7 @@ DETAIL:  distribution column value: 1
 (5 rows)
 
 -- Citus does not qualify this as a fast-path because
--- dist_key = func() 
+-- dist_key = func()
 SELECT *
 	FROM articles_hash
 	WHERE 1 = abs(author_id);
@@ -856,7 +853,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
 (5 rows)
 
 -- Citus does not qualify this as a fast-path because
--- dist_key = func() 
+-- dist_key = func()
 SELECT *
 	FROM articles_hash
 	WHERE author_id = abs(author_id - 2);
@@ -1058,7 +1055,7 @@ DETAIL:  distribution column value: 1
 (3 rows)
 
 -- window functions are supported with fast-path router plannable
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash
 	WHERE author_id = 5;
 DEBUG:  Distributed planning for a fast-path router query
@@ -1074,7 +1071,7 @@ DETAIL:  distribution column value: 5
  aminate  | aruru    |      11389
 (5 rows)
 
-SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count 
+SELECT LAG(title, 1) over (ORDER BY word_count) prev, title, word_count
 	FROM articles_hash
 	WHERE author_id = 5
 	ORDER BY word_count DESC;
@@ -1123,8 +1120,8 @@ DETAIL:  distribution column value: 1
  41 |      11814 | 7178.8000000000000000
 (5 rows)
 
-SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)  
-	FROM articles_hash 
+SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)
+	FROM articles_hash
 	WHERE author_id = 1;
 DEBUG:  Distributed planning for a fast-path router query
 DEBUG:  Creating router plan
@@ -1142,7 +1139,7 @@ DETAIL:  distribution column value: 1
 -- some more tests on complex target lists
 SELECT DISTINCT ON (author_id, id) author_id, id,
 	MIN(id) over (order by avg(word_count)) * AVG(id * 5.2 + (1.0/max(word_count))) over (order by max(word_count)) as t1,
-	count(*) FILTER (WHERE title LIKE 'al%') as cnt_with_filter, 
+	count(*) FILTER (WHERE title LIKE 'al%') as cnt_with_filter,
 	count(*) FILTER (WHERE '0300030' LIKE '%3%') as cnt_with_filter_2,
 	avg(case when id > 2 then char_length(word_count::text) * (id * strpos(word_count::text, '1')) end) as case_cnt,
 	COALESCE(strpos(avg(word_count)::text, '1'), 20)
@@ -1165,7 +1162,7 @@ DETAIL:  distribution column value: 1
 (5 rows)
 
 -- where false queries are router plannable but not fast-path
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE false;
 DEBUG:  Creating router plan
@@ -1175,7 +1172,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- fast-path with false
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE author_id = 1 and false;
 DEBUG:  Distributed planning for a fast-path router query
@@ -1186,7 +1183,7 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- fast-path with false
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE author_id = 1 and 1=0;
 DEBUG:  Distributed planning for a fast-path router query
@@ -1197,7 +1194,7 @@ DETAIL:  distribution column value: 1
 ----+-----------+-------+------------
 (0 rows)
 
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE null and author_id = 1;
 DEBUG:  Distributed planning for a fast-path router query
@@ -1209,7 +1206,7 @@ DEBUG:  Plan is router executable
 
 -- we cannot qualify dist_key = X operator Y via
 -- fast-path planning
-SELECT * 
+SELECT *
 	FROM articles_hash
 	WHERE author_id = 1 + 1;
 DEBUG:  Creating router plan
@@ -1226,7 +1223,7 @@ DETAIL:  distribution column value: 2
 
 -- where false with immutable function returning false
 -- goes through fast-path
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id = 10 and int4eq(1, 2);
 DEBUG:  Distributed planning for a fast-path router query
@@ -1240,7 +1237,7 @@ DETAIL:  distribution column value: 10
 -- partition_column is null clause does not prune out any shards,
 -- all shards remain after shard pruning, not router plannable
 -- not fast-path router either
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id is null;
 DEBUG:  Router planner cannot handle multi-shard select queries
@@ -1251,7 +1248,7 @@ DEBUG:  Router planner cannot handle multi-shard select queries
 -- partition_column equals to null clause prunes out all shards
 -- no shards after shard pruning, router plannable
 -- not fast-path router either
-SELECT * 
+SELECT *
 	FROM articles_hash a
 	WHERE a.author_id = null;
 DEBUG:  Creating router plan
@@ -1295,8 +1292,8 @@ DEBUG:  Plan is router executable
 (0 rows)
 
 -- window functions with where false
-SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)  
-	FROM articles_hash 
+SELECT word_count, rank() OVER (PARTITION BY author_id ORDER BY word_count)
+	FROM articles_hash
 	WHERE author_id = 1 and 1=0;
 DEBUG:  Distributed planning for a fast-path router query
 DEBUG:  Creating router plan
@@ -1331,7 +1328,7 @@ END;
 $$LANGUAGE plpgsql;
 SET client_min_messages TO ERROR;
 \set VERBOSITY terse
--- fast path router plannable, but errors 
+-- fast path router plannable, but errors
 SELECT raise_failed_execution_f_router($$
 	SELECT * FROM articles_hash
 		WHERE
@@ -1350,7 +1347,7 @@ SET client_min_messages TO 'DEBUG2';
 -- complex query hitting a single shard and a fast-path
 SELECT
 	count(DISTINCT CASE
-			WHEN 
+			WHEN
 				word_count > 100
 			THEN
 				id
@@ -1414,7 +1411,7 @@ END;
 WARNING:  there is no transaction in progress
 -- cursor queries are fast-path router plannable
 BEGIN;
-DECLARE test_cursor CURSOR FOR 
+DECLARE test_cursor CURSOR FOR
 	SELECT *
 		FROM articles_hash
 		WHERE author_id = 1
@@ -1465,7 +1462,6 @@ DETAIL:  distribution column value: 1
 21	1	arcading	5890
 31	1	athwartships	7271
 41	1	aznavour	11814
-	
 -- table creation queries inside can be fast-path router plannable
 CREATE TEMP TABLE temp_articles_hash as
 	SELECT *
@@ -1659,8 +1655,8 @@ BEGIN
 	return max_id;
 END;
 $$ LANGUAGE plpgsql;
--- we don't want too many details. though we're omitting 
--- "DETAIL:  distribution column value:", we see it acceptable 
+-- we don't want too many details. though we're omitting
+-- "DETAIL:  distribution column value:", we see it acceptable
 -- since the query results verifies the correctness
 \set VERBOSITY terse
 SELECT author_articles_max_id();
@@ -1936,7 +1932,7 @@ DEBUG:  Plan is router executable
 -- insert .. select via coordinator could also
 -- use fast-path queries
 PREPARE insert_sel(int, int) AS
-INSERT INTO articles_hash 
+INSERT INTO articles_hash
 	SELECT * FROM articles_hash WHERE author_id = $2 AND word_count = $1 OFFSET 0;
 EXECUTE insert_sel(1,1);
 DEBUG:  OFFSET clauses are not allowed in distributed INSERT ... SELECT queries
@@ -1983,10 +1979,10 @@ DETAIL:  distribution column value: 1
 -- one final interesting preperad statement
 -- where one of the filters is on the target list
 PREPARE fast_path_agg_filter(int, int) AS
-	SELECT 
-		count(*) FILTER (WHERE word_count=$1) 
-	FROM 
-		articles_hash 
+	SELECT
+		count(*) FILTER (WHERE word_count=$1)
+	FROM
+		articles_hash
 	WHERE author_id = $2;
 EXECUTE fast_path_agg_filter(1,1);
 DEBUG:  Distributed planning for a fast-path router query
@@ -2050,7 +2046,7 @@ DETAIL:  distribution column value: 6
 (1 row)
 
 -- views internally become subqueries, so not fast-path router query
-CREATE VIEW test_view AS 
+CREATE VIEW test_view AS
 	SELECT * FROM articles_hash WHERE author_id = 1;
 SELECT * FROM test_view;
 DEBUG:  Creating router plan
@@ -2131,7 +2127,7 @@ CREATE TABLE collections_list (
 	collection_id integer,
 	value numeric
 ) PARTITION BY LIST (collection_id );
-CREATE TABLE collections_list_1 
+CREATE TABLE collections_list_1
 	PARTITION OF collections_list (key, ts, collection_id, value)
 	FOR VALUES IN ( 1 );
 CREATE TABLE collections_list_2

src/test/regress/expected/multi_select_distinct.out

@@ -10,7 +10,7 @@ SELECT DISTINCT l_orderkey, now() FROM lineitem_hash_part LIMIT 0;
 ------------+-----
 (0 rows)
 
-SELECT DISTINCT l_partkey, 1 + (random() * 0)::int FROM lineitem_hash_part ORDER BY 1 DESC LIMIT 3; 
+SELECT DISTINCT l_partkey, 1 + (random() * 0)::int FROM lineitem_hash_part ORDER BY 1 DESC LIMIT 3;
  l_partkey | ?column? 
 -----------+----------
     199973 |        1
@@ -152,9 +152,9 @@ SELECT DISTINCT l_shipmode FROM lineitem_hash_part ORDER BY 1 DESC;
  AIR       
 (7 rows)
 
--- distinct with multiple columns 
+-- distinct with multiple columns
 SELECT DISTINCT l_orderkey, o_orderdate
-	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 	WHERE l_orderkey < 10
 	ORDER BY l_orderkey;
  l_orderkey | o_orderdate 
@@ -199,7 +199,6 @@ SELECT DISTINCT l_orderkey, count(*)
         197 |     6
 (19 rows)
 
-	
 -- explain the query to see actual plan
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT l_orderkey, count(*)
@@ -272,8 +271,7 @@ SELECT DISTINCT count(*)
      4
 (4 rows)
 
-	
--- explain the query to see actual plan. We expect to see Aggregate node having 
+-- explain the query to see actual plan. We expect to see Aggregate node having
 -- group by key on count(*) column, since columns in the Group By doesn't guarantee
 -- the uniqueness of the result.
 EXPLAIN (COSTS FALSE)
@@ -350,7 +348,6 @@ SELECT DISTINCT l_suppkey, count(*)
         14 |     1
 (10 rows)
 
-	
 -- explain the query to see actual plan. Similar to the explain of the query above.
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT l_suppkey, count(*)
@@ -377,7 +374,7 @@ EXPLAIN (COSTS FALSE)
                                        ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (15 rows)
 
--- check the plan if the hash aggreate is disabled. Similar to the explain of 
+-- check the plan if the hash aggreate is disabled. Similar to the explain of
 -- the query above.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
@@ -430,7 +427,6 @@ SELECT DISTINCT l_suppkey, avg(l_partkey)
         12 | 17510.0000000000000000
 (10 rows)
 
-	
 -- explain the query to see actual plan. Similar to the explain of the query above.
 -- Only aggregate functions will be changed.
 EXPLAIN (COSTS FALSE)
@@ -510,7 +506,6 @@ SELECT DISTINCT ON (l_suppkey) avg(l_partkey)
      77506.000000000000
 (10 rows)
 
-	
 -- explain the query to see actual plan. We expect to see sort+unique to handle
 -- distinct on.
 EXPLAIN (COSTS FALSE)
@@ -537,7 +532,7 @@ EXPLAIN (COSTS FALSE)
                                        ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (14 rows)
 
--- check the plan if the hash aggreate is disabled. We expect to see sort+unique to 
+-- check the plan if the hash aggreate is disabled. We expect to see sort+unique to
 -- handle distinct on.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
@@ -587,7 +582,6 @@ SELECT DISTINCT avg(ceil(l_partkey / 2))
  122
 (10 rows)
 
-	
 -- explain the query to see actual plan
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT avg(ceil(l_partkey / 2))
@@ -622,7 +616,7 @@ EXPLAIN (COSTS FALSE)
 		FROM lineitem_hash_part
 		GROUP BY l_suppkey, l_linenumber
 		ORDER BY 1
-		LIMIT 10;	
+		LIMIT 10;
                                                    QUERY PLAN                                                   
 ----------------------------------------------------------------------------------------------------------------
  Limit
@@ -645,11 +639,10 @@ EXPLAIN (COSTS FALSE)
                                                    ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (18 rows)
 
-		
-SET enable_hashagg TO on;		
+SET enable_hashagg TO on;
 -- expression among aggregations.
-SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis 
-	FROM lineitem_hash_part 
+SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis
+	FROM lineitem_hash_part
 	GROUP BY l_suppkey, l_linenumber
 	ORDER BY 1
 	LIMIT 10;
@@ -667,11 +660,10 @@ SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis
   15
 (10 rows)
 
-	
 -- explain the query to see actual plan
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis
+		FROM lineitem_hash_part
 		GROUP BY l_suppkey, l_linenumber
 		ORDER BY 1
 		LIMIT 10;
@@ -698,11 +690,11 @@ EXPLAIN (COSTS FALSE)
 -- to a bug right now, expectation must be corrected after fixing it.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT sum(l_suppkey) + count(l_partkey) AS dis
+		FROM lineitem_hash_part
 		GROUP BY l_suppkey, l_linenumber
 		ORDER BY 1
-		LIMIT 10;	
+		LIMIT 10;
                                                                     QUERY PLAN                                                                    
 --------------------------------------------------------------------------------------------------------------------------------------------------
  Limit
@@ -725,13 +717,11 @@ EXPLAIN (COSTS FALSE)
                                                    ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (18 rows)
 
-		
-SET enable_hashagg TO on;		
-		
+SET enable_hashagg TO on;
 -- distinct on all columns, note Group By columns guarantees uniqueness of the
--- result list.		
-SELECT DISTINCT * 
-	FROM lineitem_hash_part 
+-- result list.
+SELECT DISTINCT *
+	FROM lineitem_hash_part
 	GROUP BY 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16
 	ORDER BY 1,2
 	LIMIT 10;
@@ -749,12 +739,11 @@ SELECT DISTINCT *
           3 |     29380 |      1883 |            4 |       2.00 |         2618.76 |       0.01 |  0.06 | A            | F            | 12-04-1993 | 01-07-1994   | 01-01-1994    | NONE                      | TRUCK      | y. fluffily pending d
 (10 rows)
 
-	
 -- explain the query to see actual plan. We expect to see only one aggregation
 -- node since group by columns guarantees the uniqueness.
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT * 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT *
+		FROM lineitem_hash_part
 		GROUP BY 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16
 		ORDER BY 1,2
 		LIMIT 10;
@@ -779,15 +768,15 @@ EXPLAIN (COSTS FALSE)
                                                    ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (17 rows)
 
--- check the plan if the hash aggreate is disabled. We expect to see only one 
+-- check the plan if the hash aggreate is disabled. We expect to see only one
 -- aggregation node since group by columns guarantees the uniqueness.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT * 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT *
+		FROM lineitem_hash_part
 		GROUP BY 1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16
 		ORDER BY 1,2
-		LIMIT 10;	
+		LIMIT 10;
                                                                                                                                                                                                                 QUERY PLAN                                                                                                                                                                                                                
 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
  Limit
@@ -811,7 +800,6 @@ EXPLAIN (COSTS FALSE)
                                                          ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (19 rows)
 
-		
 SET enable_hashagg TO on;
 -- distinct on count distinct
 SELECT DISTINCT count(DISTINCT l_partkey), count(DISTINCT l_shipmode)
@@ -847,7 +835,6 @@ SELECT DISTINCT count(DISTINCT l_partkey), count(DISTINCT l_shipmode)
      7 |     7
 (25 rows)
 
-	
 -- explain the query to see actual plan. We expect to see aggregation plan for
 -- the outer distinct.
 EXPLAIN (COSTS FALSE)
@@ -875,7 +862,7 @@ EXPLAIN (COSTS FALSE)
                                        ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (16 rows)
 
--- check the plan if the hash aggreate is disabled. We expect to see sort + unique 
+-- check the plan if the hash aggreate is disabled. We expect to see sort + unique
 -- plans for the outer distinct.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
@@ -906,11 +893,10 @@ EXPLAIN (COSTS FALSE)
                                                    ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (19 rows)
 
-		
 SET enable_hashagg TO on;
 -- distinct on aggregation with filter and expression
-SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count 
-	FROM lineitem_hash_part 
+SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count
+	FROM lineitem_hash_part
 	GROUP BY l_suppkey
 	ORDER BY 1;
  count 
@@ -922,11 +908,10 @@ SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2)
      4
 (5 rows)
 
-	
 -- explain the query to see actual plan
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count
+		FROM lineitem_hash_part
 		GROUP BY l_suppkey
 		ORDER BY 1;
                                                       QUERY PLAN                                                       
@@ -950,8 +935,8 @@ EXPLAIN (COSTS FALSE)
 -- check the plan if the hash aggreate is disabled
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
-	SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count 
-		FROM lineitem_hash_part 
+	SELECT DISTINCT ceil(count(case when l_partkey > 100000 THEN 1 ELSE 0 END) / 2) AS count
+		FROM lineitem_hash_part
 		GROUP BY l_suppkey
 		ORDER BY 1;
                                                           QUERY PLAN                                                          
@@ -975,13 +960,11 @@ EXPLAIN (COSTS FALSE)
                                              ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (17 rows)
 
-		
 SET enable_hashagg TO on;
-	
--- explain the query to see actual plan with array_agg aggregation.  
+-- explain the query to see actual plan with array_agg aggregation.
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT array_agg(l_linenumber), array_length(array_agg(l_linenumber), 1)
-		FROM lineitem_hash_part 
+		FROM lineitem_hash_part
 		GROUP BY l_orderkey
 		ORDER BY 2
 		LIMIT 15;
@@ -1006,11 +989,11 @@ EXPLAIN (COSTS FALSE)
                                              ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (17 rows)
 
--- check the plan if the hash aggreate is disabled. 
+-- check the plan if the hash aggreate is disabled.
 SET enable_hashagg TO off;
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT array_agg(l_linenumber), array_length(array_agg(l_linenumber), 1)
-		FROM lineitem_hash_part 
+		FROM lineitem_hash_part
 		GROUP BY l_orderkey
 		ORDER BY 2
 		LIMIT 15;
@@ -1038,13 +1021,12 @@ EXPLAIN (COSTS FALSE)
                                                          ->  Seq Scan on lineitem_hash_part_360041 lineitem_hash_part
 (20 rows)
 
-		
 SET enable_hashagg TO on;
 -- distinct on non-partition column with aggregate
 -- this is the same as non-distinct version due to group by
 SELECT DISTINCT l_partkey, count(*)
 	FROM lineitem_hash_part
-	GROUP BY 1 
+	GROUP BY 1
 	HAVING count(*) > 2
 	ORDER BY 1;
  l_partkey | count 
@@ -1062,12 +1044,11 @@ SELECT DISTINCT l_partkey, count(*)
     199146 |     3
 (11 rows)
 
-	
 -- explain the query to see actual plan
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT l_partkey, count(*)
 		FROM lineitem_hash_part
-		GROUP BY 1 
+		GROUP BY 1
 		HAVING count(*) > 2
 		ORDER BY 1;
                                              QUERY PLAN                                             
@@ -1091,7 +1072,7 @@ EXPLAIN (COSTS FALSE)
 SELECT DISTINCT l_partkey, avg(l_linenumber)
 	FROM lineitem_hash_part
 	WHERE l_partkey < 500
-	GROUP BY 1 
+	GROUP BY 1
 	HAVING avg(l_linenumber) > 2
 	ORDER BY 1;
  l_partkey |        avg         
@@ -1137,7 +1118,6 @@ SELECT DISTINCT l_partkey, l_suppkey
     197921 |       441
 (15 rows)
 
-	
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT l_partkey, l_suppkey
 		FROM lineitem_hash_part
@@ -1180,7 +1160,6 @@ SELECT DISTINCT ON (l_orderkey) l_orderkey, l_partkey, l_suppkey
          34 |     88362 |       871
 (10 rows)
 
-	
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT ON (l_orderkey) l_orderkey, l_partkey, l_suppkey
 		FROM lineitem_hash_part
@@ -1261,7 +1240,7 @@ EXPLAIN (COSTS FALSE)
 -- distinct on with joins
 -- each customer's first order key
 SELECT DISTINCT ON (o_custkey) o_custkey, l_orderkey
-	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 	WHERE o_custkey < 15
 	ORDER BY 1,2;
  o_custkey | l_orderkey 
@@ -1281,7 +1260,7 @@ SELECT DISTINCT ON (o_custkey) o_custkey, l_orderkey
 SELECT coordinator_plan($Q$
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT ON (o_custkey) o_custkey, l_orderkey
-		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 		WHERE o_custkey < 15
 		ORDER BY 1,2;
 $Q$);
@@ -1299,7 +1278,7 @@ $Q$);
 SELECT coordinator_plan($Q$
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT ON (o_custkey) o_custkey, l_orderkey
-		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 		WHERE o_custkey < 15;
 $Q$);
              coordinator_plan             
@@ -1313,7 +1292,7 @@ $Q$);
 
 -- each customer's each order's first l_partkey
 SELECT DISTINCT ON (o_custkey, l_orderkey) o_custkey, l_orderkey, l_linenumber, l_partkey
-	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 	WHERE o_custkey < 20
 	ORDER BY 1,2,3;
  o_custkey | l_orderkey | l_linenumber | l_partkey 
@@ -1360,7 +1339,7 @@ SELECT DISTINCT ON (o_custkey, l_orderkey) o_custkey, l_orderkey, l_linenumber,
 SELECT coordinator_plan($Q$
 EXPLAIN (COSTS FALSE)
 	SELECT DISTINCT ON (o_custkey, l_orderkey) o_custkey, l_orderkey, l_linenumber, l_partkey
-		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+		FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 		WHERE o_custkey < 20;
 $Q$);
                         coordinator_plan                         
@@ -1374,7 +1353,7 @@ $Q$);
 
 -- each customer's each order's last l_partkey
 SELECT DISTINCT ON (o_custkey, l_orderkey) o_custkey, l_orderkey, l_linenumber, l_partkey
-	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey) 
+	FROM lineitem_hash_part JOIN orders_hash_part ON (l_orderkey = o_orderkey)
 	WHERE o_custkey < 15
 	ORDER BY 1,2,3 DESC;
  o_custkey | l_orderkey | l_linenumber | l_partkey 

src/test/regress/expected/multi_subquery_complex_reference_clause.out

@@ -374,7 +374,6 @@ DEBUG:  Plan is router executable
      5
 (1 row)
 
- 
 -- union involving reference table and distributed table subqueries
 -- is supported with pulling data to coordinator
 SELECT * FROM
@@ -1512,7 +1511,7 @@ ORDER BY 1 LIMIT 3;
 
 -- should be able to pushdown since one of the subqueries has distinct on reference tables
 -- and there is only reference table in that subquery
-SELECT 
+SELECT
   distinct_users, event_type, time
 FROM
 (SELECT user_id, time, event_type FROM events_table) as events_dist INNER JOIN
@@ -1530,11 +1529,11 @@ OFFSET 0;
 (5 rows)
 
 -- the same query wuth multiple reference tables in the subquery
-SELECT 
+SELECT
   distinct_users, event_type, time
 FROM
 (SELECT user_id, time, event_type FROM events_table) as events_dist INNER JOIN
-(SELECT DISTINCT users_reference_table.user_id as distinct_users FROM users_reference_table, events_reference_table 
+(SELECT DISTINCT users_reference_table.user_id as distinct_users FROM users_reference_table, events_reference_table
  WHERE events_reference_table.user_id =  users_reference_table.user_id AND events_reference_table.event_type IN (1,2,3,4)) users_ref
 ON (events_dist.user_id = users_ref.distinct_users)
 ORDER BY time DESC
@@ -1550,7 +1549,7 @@ OFFSET 0;
 (5 rows)
 
 -- similar query as the above, but with group bys
-SELECT 
+SELECT
   distinct_users, event_type, time
 FROM
 (SELECT user_id, time, event_type FROM events_table) as events_dist INNER JOIN
@@ -1587,7 +1586,7 @@ SELECT * FROM
 ---------
 (0 rows)
 
--- similiar to the above examples, this time there is a subquery 
+-- similiar to the above examples, this time there is a subquery
 -- whose output is not in the DISTINCT clause
 SELECT * FROM
 (
@@ -1608,8 +1607,8 @@ ORDER BY 1;
 SELECT * FROM
 (
   SELECT DISTINCT users_reference_table.user_id, us_events.user_id FROM users_reference_table, (SELECT user_id, random() FROM events_table WHERE event_type IN (2,3)) as us_events WHERE users_reference_table.user_id = us_events.user_id
-) as foo 
-ORDER BY 1 DESC 
+) as foo
+ORDER BY 1 DESC
 LIMIT 4;
  user_id | user_id 
 ---------+---------
@@ -1624,15 +1623,15 @@ LIMIT 4;
 -- is disabled
 SELECT * FROM
 (
-  SELECT 
-    DISTINCT users_reference_table.user_id, us_events.value_4 
-  FROM 
-    users_reference_table, 
-    (SELECT user_id, value_4, random() FROM events_table WHERE event_type IN (2,3)) as us_events 
-  WHERE 
+  SELECT
+    DISTINCT users_reference_table.user_id, us_events.value_4
+  FROM
+    users_reference_table,
+    (SELECT user_id, value_4, random() FROM events_table WHERE event_type IN (2,3)) as us_events
+  WHERE
     users_reference_table.user_id = us_events.user_id
-) as foo 
-ORDER BY 1 DESC 
+) as foo
+ORDER BY 1 DESC
 LIMIT 4;
  user_id | value_4 
 ---------+---------
@@ -1644,7 +1643,6 @@ LIMIT 4;
 
 -- test the read_intermediate_result() for GROUP BYs
 BEGIN;
- 
 SELECT broadcast_intermediate_result('squares', 'SELECT s, s*s FROM generate_series(1,200) s');
  broadcast_intermediate_result 
 -------------------------------
@@ -1652,15 +1650,15 @@ SELECT broadcast_intermediate_result('squares', 'SELECT s, s*s FROM generate_ser
 (1 row)
 
 -- single appereance of read_intermediate_result
-SELECT 
-  DISTINCT user_id 
-FROM 
-  users_table 
-JOIN 
-(SELECT 
-  max(res.val) as mx 
-FROM 
-  read_intermediate_result('squares', 'binary') AS res (val int, val_square int) 
+SELECT
+  DISTINCT user_id
+FROM
+  users_table
+JOIN
+(SELECT
+  max(res.val) as mx
+FROM
+  read_intermediate_result('squares', 'binary') AS res (val int, val_square int)
 GROUP BY res.val_square) squares
  ON (mx = user_id)
 ORDER BY 1
@@ -1692,15 +1690,15 @@ LIMIT 5;
 (5 rows)
 
 -- single appereance of read_intermediate_result but inside a subquery
-SELECT 
-  DISTINCT user_id 
-FROM 
-  users_table 
+SELECT
+  DISTINCT user_id
+FROM
+  users_table
 JOIN (
-  SELECT *,random() FROM (SELECT 
-    max(res.val) as mx 
-  FROM 
-      (SELECT val, val_square FROM read_intermediate_result('squares', 'binary') AS res (val int, val_square int)) res 
+  SELECT *,random() FROM (SELECT
+    max(res.val) as mx
+  FROM
+      (SELECT val, val_square FROM read_intermediate_result('squares', 'binary') AS res (val int, val_square int)) res
   GROUP BY res.val_square) foo)
 squares
  ON (mx = user_id)
@@ -1716,16 +1714,16 @@ LIMIT 5;
 (5 rows)
 
 -- multiple read_intermediate_results in the same subquery is OK
-SELECT 
-  DISTINCT user_id 
-FROM 
-  users_table 
-JOIN 
-(SELECT 
-  max(res.val) as mx 
-FROM 
+SELECT
+  DISTINCT user_id
+FROM
+  users_table
+JOIN
+(SELECT
+  max(res.val) as mx
+FROM
   read_intermediate_result('squares', 'binary') AS res (val int, val_square int),
-  read_intermediate_result('squares', 'binary') AS res2 (val int, val_square int) 
+  read_intermediate_result('squares', 'binary') AS res2 (val int, val_square int)
 WHERE res.val = res2.val_square
 GROUP BY res2.val_square) squares
  ON (mx = user_id)
@@ -1738,19 +1736,19 @@ LIMIT 5;
 (2 rows)
 
 -- mixed recurring tuples should be supported
-SELECT 
-  DISTINCT user_id 
-FROM 
-  users_table 
-JOIN 
-(SELECT 
-  max(res.val) as mx 
-FROM 
-  read_intermediate_result('squares', 'binary') AS res (val int, val_square int), 
+SELECT
+  DISTINCT user_id
+FROM
+  users_table
+JOIN
+(SELECT
+  max(res.val) as mx
+FROM
+  read_intermediate_result('squares', 'binary') AS res (val int, val_square int),
   generate_series(0, 10) i
   WHERE
   res.val = i
-  GROUP BY 
+  GROUP BY
     i) squares
  ON (mx = user_id)
 ORDER BY 1
@@ -1766,16 +1764,16 @@ LIMIT 5;
 
 -- should recursively plan since
 -- there are no columns on the GROUP BY from the distributed table
-SELECT 
-  DISTINCT user_id 
-FROM 
+SELECT
+  DISTINCT user_id
+FROM
   users_reference_table
-JOIN 
-  (SELECT 
-    max(val_square) as mx 
-  FROM 
-    read_intermediate_result('squares', 'binary') AS res (val int, val_square int), events_table 
-  WHERE 
+JOIN
+  (SELECT
+    max(val_square) as mx
+  FROM
+    read_intermediate_result('squares', 'binary') AS res (val int, val_square int), events_table
+  WHERE
     events_table.user_id = res.val GROUP BY res.val) squares
  ON (mx = user_id)
 ORDER BY 1
@@ -1788,14 +1786,14 @@ LIMIT 5;
 
 ROLLBACK;
 -- should work since we're using an immutable function as recurring tuple
-SELECT 
-  DISTINCT user_id 
-FROM 
-  users_table 
-JOIN 
-(SELECT 
-  max(i+5)as mx 
-FROM 
+SELECT
+  DISTINCT user_id
+FROM
+  users_table
+JOIN
+(SELECT
+  max(i+5)as mx
+FROM
   generate_series(0, 10) as i GROUP BY i) squares
  ON (mx = user_id)
 ORDER BY 1
@@ -1806,20 +1804,20 @@ LIMIT 5;
        6
 (2 rows)
 
--- should recursively plan since we're 
+-- should recursively plan since we're
 -- using an immutable function as recurring tuple
--- along with a distributed table, where GROUP BY is 
+-- along with a distributed table, where GROUP BY is
 -- on the recurring tuple
-SELECT 
-  DISTINCT user_id 
-FROM 
+SELECT
+  DISTINCT user_id
+FROM
   users_reference_table
-JOIN 
-  (SELECT 
-    max(i+5)as mx 
-  FROM 
-     generate_series(0, 10) as i, events_table 
-  WHERE 
+JOIN
+  (SELECT
+    max(i+5)as mx
+  FROM
+     generate_series(0, 10) as i, events_table
+  WHERE
     events_table.user_id = i GROUP BY i) squares
  ON (mx = user_id)
 ORDER BY 1

src/test/regress/expected/multi_upgrade_reference_table.out

@@ -228,7 +228,6 @@ ORDER BY shardid;
  1360009 | t
 (1 row)
 
-    
 DROP TABLE upgrade_reference_table_append;
 -- test valid cases, shard exists at one worker
 CREATE TABLE upgrade_reference_table_one_worker(column1 int);
@@ -341,7 +340,6 @@ ORDER BY shardid;
  1360010 | t
 (1 row)
 
-    
 DROP TABLE upgrade_reference_table_one_worker;
 -- test valid cases, shard exists at both workers but one is unhealthy
 SET citus.shard_replication_factor TO 2;
@@ -458,7 +456,6 @@ ORDER BY shardid;
  1360011 | t
 (1 row)
 
-    
 DROP TABLE upgrade_reference_table_one_unhealthy;
 -- test valid cases, shard exists at both workers and both are healthy
 CREATE TABLE upgrade_reference_table_both_healthy(column1 int);
@@ -570,7 +567,6 @@ ORDER BY shardid;
  1360012 | t
 (1 row)
 
-    
 DROP TABLE upgrade_reference_table_both_healthy;
 -- test valid cases, do it in transaction and ROLLBACK
 SET citus.shard_replication_factor TO 1;
@@ -686,7 +682,6 @@ ORDER BY shardid;
  1360013 | f
 (1 row)
 
-     
 DROP TABLE upgrade_reference_table_transaction_rollback;
 -- test valid cases, do it in transaction and COMMIT
 SET citus.shard_replication_factor TO 1;
@@ -872,11 +867,9 @@ ORDER BY shardid;
  1360015
 (1 row)
 
-     
 SELECT upgrade_to_reference_table('upgrade_reference_table_mx');
 ERROR:  cannot upgrade to reference table
 DETAIL:  Upgrade is only supported for statement-based replicated tables but "upgrade_reference_table_mx" is streaming replicated
-     
 -- situation after upgrade_reference_table
 SELECT
     partmethod, (partkey IS NULL) as partkeyisnull, colocationid, repmodel
@@ -937,10 +930,9 @@ SELECT create_distributed_table('upgrade_reference_table_mx', 'column1');
  
 (1 row)
 
-UPDATE pg_dist_shard_placement SET shardstate = 3 
-WHERE nodeport = :worker_2_port AND 
+UPDATE pg_dist_shard_placement SET shardstate = 3
+WHERE nodeport = :worker_2_port AND
 	shardid IN (SELECT shardid FROM pg_dist_shard WHERE logicalrelid='upgrade_reference_table_mx'::regclass);
-	
 SELECT start_metadata_sync_to_node('localhost', :worker_1_port);
  start_metadata_sync_to_node 
 -----------------------------
@@ -995,7 +987,6 @@ ORDER BY shardid;
  1360016
 (1 row)
 
-     
 SET client_min_messages TO WARNING;
 SELECT upgrade_to_reference_table('upgrade_reference_table_mx');
  upgrade_to_reference_table 
@@ -1003,7 +994,6 @@ SELECT upgrade_to_reference_table('upgrade_reference_table_mx');
  
 (1 row)
 
-     
 -- situation after upgrade_reference_table
 SELECT
     partmethod, (partkey IS NULL) as partkeyisnull, colocationid, repmodel
@@ -1052,7 +1042,6 @@ ORDER BY shardid;
  1360016 | t
 (1 row)
 
-     
 -- situation on metadata worker
 \c - - - :worker_1_port
 SELECT
@@ -1091,7 +1080,6 @@ ORDER BY shardid;
  1360016 | t
 (1 row)
 
-     
 \c - - - :master_port
 DROP TABLE upgrade_reference_table_mx;
 SELECT stop_metadata_sync_to_node('localhost', :worker_1_port);

src/test/regress/expected/propagate_extension_commands.out

@@ -51,7 +51,6 @@ BEGIN;
  
 (1 row)
 
-	
 	-- we can even run queries (sequentially) over the distributed table
 	SELECT * FROM dist_table;
  key | value 
@@ -103,7 +102,7 @@ SELECT run_command_on_workers($$SELECT extversion FROM pg_extension WHERE extnam
 
 -- now, update to a newer version
 ALTER EXTENSION isn UPDATE TO '1.2';
--- show that ALTER EXTENSION is propagated 
+-- show that ALTER EXTENSION is propagated
 SELECT run_command_on_workers($$SELECT extversion FROM pg_extension WHERE extname = 'isn'$$);
  run_command_on_workers  
 -------------------------
@@ -145,8 +144,8 @@ DROP EXTENSION isn CASCADE;
 -- restore client_min_messages after DROP EXTENSION
 RESET client_min_messages;
 -- now make sure that the reference tables depending on an extension can be succesfully created.
--- we should also ensure that we replicate this reference table (and hence the extension) 
--- to new nodes after calling master_activate_node. 
+-- we should also ensure that we replicate this reference table (and hence the extension)
+-- to new nodes after calling master_activate_node.
 -- now, first drop seg and existing objects before next test
 SET client_min_messages TO WARNING;
 DROP EXTENSION seg CASCADE;
@@ -207,7 +206,7 @@ SELECT run_command_on_workers($$SELECT extversion FROM pg_extension WHERE extnam
 (2 rows)
 
 -- and similarly check for the reference table
-select count(*) from pg_dist_partition where partmethod='n' and logicalrelid='ref_table_2'::regclass; 
+select count(*) from pg_dist_partition where partmethod='n' and logicalrelid='ref_table_2'::regclass;
  count 
 -------
      1
@@ -241,7 +240,7 @@ SELECT run_command_on_workers($$SELECT count(*) FROM pg_extension WHERE extname
 (2 rows)
 
 -- give a notice for the following commands saying that it is not
--- propagated to the workers. the user should run it manually on the workers 
+-- propagated to the workers. the user should run it manually on the workers
 CREATE TABLE t1 (A int);
 CREATE VIEW v1 AS select * from t1;
 ALTER EXTENSION seg ADD VIEW v1;
@@ -346,8 +345,8 @@ BEGIN;
 	SET citus.shard_replication_factor TO 1;
 	CREATE EXTENSION seg;
 	CREATE EXTENSION isn;
-	CREATE TYPE test_type AS (a int, b seg); 
-	CREATE TYPE test_type_2 AS (a int, b test_type); 
+	CREATE TYPE test_type AS (a int, b seg);
+	CREATE TYPE test_type_2 AS (a int, b test_type);
 	CREATE TABLE t2 (a int, b test_type_2, c issn);
 	SELECT create_distributed_table('t2', 'a');
  create_distributed_table 
@@ -355,8 +354,7 @@ BEGIN;
  
 (1 row)
 
-	
-	CREATE TYPE test_type_3 AS (a int, b test_type, c issn); 
+	CREATE TYPE test_type_3 AS (a int, b test_type, c issn);
 	CREATE TABLE t3 (a int, b test_type_3);
 	SELECT create_reference_table('t3');
  create_reference_table 

src/test/regress/expected/sequential_modifications.out

@@ -1,8 +1,8 @@
--- 
+--
 -- Tests sequential and parallel DDL command execution
 -- in combination with 1PC and 2PC
 -- Note: this test should not be executed in parallel with
--- other tests since we're relying on disabling 2PC recovery 
+-- other tests since we're relying on disabling 2PC recovery
 --
 -- We're also setting force_max_query_parallelization throughout
 -- the tests because the test file has the assumption that
@@ -29,7 +29,7 @@ $$
 LANGUAGE 'plpgsql' IMMUTABLE;
 -- this function simply checks the equality of the number of transactions in the
 -- pg_dist_transaction and number of shard placements for a distributed table
--- The function is useful to ensure that a single connection is opened per 
+-- The function is useful to ensure that a single connection is opened per
 -- shard placement in a distributed transaction
 CREATE OR REPLACE FUNCTION distributed_2PCs_are_equal_to_placement_count()
     RETURNS bool AS
@@ -184,7 +184,7 @@ SELECT distributed_2PCs_are_equal_to_worker_count();
  t
 (1 row)
 
--- tables with replication factor > 1 should also obey 
+-- tables with replication factor > 1 should also obey
 -- both multi_shard_commit_protocol and multi_shard_modify_mode
 SET citus.shard_replication_factor TO 2;
 CREATE TABLE test_table_rep_2 (a int);
@@ -194,7 +194,6 @@ SELECT create_distributed_table('test_table_rep_2', 'a');
  
 (1 row)
 
- 
 -- 1PC should never use 2PC with rep > 1
 SET citus.multi_shard_commit_protocol TO '1pc';
 SET citus.multi_shard_modify_mode TO 'sequential';

src/test/regress/expected/subquery_and_cte.out

@@ -26,19 +26,19 @@ WITH cte AS (
 	)
 	SELECT dist_cte.user_id FROM local_cte JOIN dist_cte ON dist_cte.user_id=local_cte.user_id
 )
-SELECT 
-	count(*) 
-FROM 
+SELECT
+	count(*)
+FROM
 	cte,
-	  (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
-     	users_table, events_table 
-     WHERE 
-     	users_table.user_id = events_table.user_id AND 
+	  (SELECT
+    	DISTINCT users_table.user_id
+     FROM
+     	users_table, events_table
+     WHERE
+     	users_table.user_id = events_table.user_id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
-     ) as foo 
+     ) as foo
 	  WHERE foo.user_id = cte.user_id;
 DEBUG:  generating subplan 3_1 for CTE cte: WITH local_cte AS (SELECT users_table_local.user_id, users_table_local."time", users_table_local.value_1, users_table_local.value_2, users_table_local.value_3, users_table_local.value_4 FROM subquery_and_ctes.users_table_local), dist_cte AS (SELECT events_table.user_id FROM subquery_and_ctes.events_table) SELECT dist_cte.user_id FROM (local_cte JOIN dist_cte ON ((dist_cte.user_id OPERATOR(pg_catalog.=) local_cte.user_id)))
 DEBUG:  generating subplan 4_1 for CTE local_cte: SELECT user_id, "time", value_1, value_2, value_3, value_4 FROM subquery_and_ctes.users_table_local
@@ -125,8 +125,8 @@ FROM cte1 WHERE dt.id = 1;
 DEBUG:  generating subplan 16_1 for CTE cte1: SELECT id, value FROM subquery_and_ctes.func() func(id, value)
 DEBUG:  Plan 16 query after replacing subqueries and CTEs: UPDATE subquery_and_ctes.dist_table dt SET value = cte1.value FROM (SELECT intermediate_result.id, intermediate_result.value FROM read_intermediate_result('16_1'::text, 'binary'::citus_copy_format) intermediate_result(id integer, value integer)) cte1 WHERE (dt.id OPERATOR(pg_catalog.=) 1)
 -- CTEs are recursively planned, and subquery foo is also recursively planned
--- final plan becomes a real-time plan since we also have events_table in the 
--- range table entries 
+-- final plan becomes a real-time plan since we also have events_table in the
+-- range table entries
 WITH cte AS (
 	WITH local_cte AS (
 		SELECT * FROM users_table_local
@@ -136,16 +136,16 @@ WITH cte AS (
 	)
 	SELECT dist_cte.user_id FROM local_cte JOIN dist_cte ON dist_cte.user_id=local_cte.user_id
 )
-SELECT 
-	count(*) 
-FROM 
+SELECT
+	count(*)
+FROM
 	cte,
-	  (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
-     	users_table, events_table 
-     WHERE 
-     	users_table.user_id = events_table.user_id AND 
+	  (SELECT
+    	DISTINCT users_table.user_id
+     FROM
+     	users_table, events_table
+     WHERE
+     	users_table.user_id = events_table.user_id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
      ) as foo, events_table
@@ -163,7 +163,7 @@ DEBUG:  Plan 17 query after replacing subqueries and CTEs: SELECT count(*) AS co
 (1 row)
 
 -- CTEs are replaced and subquery in WHERE is also replaced
--- but the query is still real-time query since users_table is in the 
+-- but the query is still real-time query since users_table is in the
 -- range table list
 WITH cte AS (
 	WITH local_cte AS (
@@ -176,7 +176,7 @@ WITH cte AS (
 )
 SELECT DISTINCT cte.user_id
 FROM users_table, cte
-WHERE 
+WHERE
 	users_table.user_id = cte.user_id AND
 	users_table.user_id IN (SELECT DISTINCT value_2 FROM users_table WHERE value_1 >= 1 AND value_1 <= 20 ORDER BY 1 LIMIT 5)
     ORDER BY 1 DESC;
@@ -208,7 +208,7 @@ WITH cte AS (
 )
 SELECT DISTINCT cte.user_id
 FROM cte
-WHERE 
+WHERE
 	cte.user_id IN (SELECT DISTINCT user_id FROM users_table WHERE value_1 >= 1 AND value_1 <= 20)
     ORDER BY 1 DESC;
 DEBUG:  generating subplan 25_1 for CTE cte: WITH local_cte AS (SELECT users_table_local.user_id, users_table_local."time", users_table_local.value_1, users_table_local.value_2, users_table_local.value_3, users_table_local.value_4 FROM subquery_and_ctes.users_table_local), dist_cte AS (SELECT events_table.user_id FROM subquery_and_ctes.events_table) SELECT dist_cte.user_id FROM (local_cte JOIN dist_cte ON ((dist_cte.user_id OPERATOR(pg_catalog.=) local_cte.user_id)))
@@ -234,12 +234,12 @@ SELECT
 FROM
     (
 	     WITH cte AS (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
 	     ) SELECT * FROM cte ORDER BY 1 DESC
      ) as foo
@@ -263,25 +263,24 @@ SELECT
 FROM
     (
 	     WITH cte AS (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
 	     ) SELECT * FROM cte ORDER BY 1 DESC
-     ) as foo, 
+     ) as foo,
     (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
-	    
-     ) as bar  
+     ) as bar
 WHERE foo.user_id = bar.user_id
 ORDER BY 1 DESC;
 DEBUG:  generating subplan 31_1 for CTE cte: SELECT DISTINCT users_table.user_id FROM subquery_and_ctes.users_table, subquery_and_ctes.events_table WHERE ((users_table.user_id OPERATOR(pg_catalog.=) events_table.user_id) AND (events_table.event_type OPERATOR(pg_catalog.=) ANY (ARRAY[1, 2, 3, 4])))
@@ -303,38 +302,37 @@ SELECT
 FROM
     (
 	     WITH cte AS (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
 	     ) SELECT * FROM cte ORDER BY 1 DESC
-     ) as foo, 
+     ) as foo,
     (
-	    SELECT 
+	    SELECT
 	    	users_table.user_id, some_events.event_type
-	     FROM 
-	     	users_table, 
+	     FROM
+	     	users_table,
 	     	(
 	     		WITH cte AS (
-			    SELECT 
+			    SELECT
 			    	event_type, users_table.user_id
-			     FROM 
-			     	users_table, events_table 
-			     WHERE 
-			     	users_table.user_id = events_table.user_id AND 
+			     FROM
+			     	users_table, events_table
+			     WHERE
+			     	users_table.user_id = events_table.user_id AND
 			     value_1 IN (1,2)
 			     ) SELECT * FROM cte ORDER BY 1 DESC
 	     	) as some_events
-	     WHERE 
-	     	users_table.user_id = some_events.user_id AND 
+	     WHERE
+	     	users_table.user_id = some_events.user_id AND
 	     event_type IN (1,2,3,4)
 	     ORDER BY 2,1
 	     LIMIT 2
-	    
-     ) as bar  
+     ) as bar
 WHERE foo.user_id = bar.user_id
 ORDER BY 1 DESC LIMIT 5;
 DEBUG:  generating subplan 33_1 for CTE cte: SELECT DISTINCT users_table.user_id FROM subquery_and_ctes.users_table, subquery_and_ctes.events_table WHERE ((users_table.user_id OPERATOR(pg_catalog.=) events_table.user_id) AND (events_table.event_type OPERATOR(pg_catalog.=) ANY (ARRAY[1, 2, 3, 4])))
@@ -347,10 +345,10 @@ DEBUG:  Plan 33 query after replacing subqueries and CTEs: SELECT DISTINCT bar.u
        1
 (1 row)
 
--- CTEs on the different parts of the query is replaced 
+-- CTEs on the different parts of the query is replaced
 -- and subquery foo is also replaced since it contains
--- DISTINCT on a non-partition key 
-SELECT * FROM 
+-- DISTINCT on a non-partition key
+SELECT * FROM
 (
 	WITH cte AS (
 		WITH local_cte AS (
@@ -363,16 +361,16 @@ SELECT * FROM
 	)
 	SELECT DISTINCT cte.user_id
 	FROM users_table, cte
-	WHERE 
+	WHERE
 		users_table.user_id = cte.user_id AND
-		users_table.user_id IN 
+		users_table.user_id IN
 			(WITH cte_in_where AS (SELECT DISTINCT value_2 FROM users_table WHERE value_1 >= 1 AND value_1 <= 20 ORDER BY 1 LIMIT 5) SELECT * FROM cte_in_where)
 	    ORDER BY 1 DESC
-	    ) as foo, 
-			events_table 
-		WHERE 
+	    ) as foo,
+			events_table
+		WHERE
 			foo.user_id = events_table.value_2
-ORDER BY 3 DESC, 2 DESC, 1 DESC 
+ORDER BY 3 DESC, 2 DESC, 1 DESC
 LIMIT 5;
 DEBUG:  generating subplan 37_1 for CTE cte: WITH local_cte AS (SELECT users_table_local.user_id, users_table_local."time", users_table_local.value_1, users_table_local.value_2, users_table_local.value_3, users_table_local.value_4 FROM subquery_and_ctes.users_table_local), dist_cte AS (SELECT events_table.user_id FROM subquery_and_ctes.events_table) SELECT dist_cte.user_id FROM (local_cte JOIN dist_cte ON ((dist_cte.user_id OPERATOR(pg_catalog.=) local_cte.user_id)))
 DEBUG:  generating subplan 38_1 for CTE local_cte: SELECT user_id, "time", value_1, value_2, value_3, value_4 FROM subquery_and_ctes.users_table_local
@@ -398,30 +396,30 @@ WITH cte AS (
 		SELECT * FROM users_table_local
 	),
 	dist_cte AS (
-		SELECT 
+		SELECT
 			user_id
-		FROM 
-			events_table, 
+		FROM
+			events_table,
 			(SELECT DISTINCT value_2 FROM users_table OFFSET 0) as foo
-		WHERE 
+		WHERE
 			events_table.user_id = foo.value_2 AND
 			events_table.user_id IN (SELECT DISTINCT value_1 FROM users_table ORDER BY 1 LIMIT 3)
 	)
 	SELECT dist_cte.user_id FROM local_cte JOIN dist_cte ON dist_cte.user_id=local_cte.user_id
 )
-SELECT 
-	count(*) 
-FROM 
+SELECT
+	count(*)
+FROM
 	cte,
-	  (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
-     	users_table, events_table 
-     WHERE 
-     	users_table.user_id = events_table.user_id AND 
+	  (SELECT
+    	DISTINCT users_table.user_id
+     FROM
+     	users_table, events_table
+     WHERE
+     	users_table.user_id = events_table.user_id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
-     ) as foo 
+     ) as foo
 	  WHERE foo.user_id = cte.user_id;
 DEBUG:  generating subplan 42_1 for CTE cte: WITH local_cte AS (SELECT users_table_local.user_id, users_table_local."time", users_table_local.value_1, users_table_local.value_2, users_table_local.value_3, users_table_local.value_4 FROM subquery_and_ctes.users_table_local), dist_cte AS (SELECT events_table.user_id FROM subquery_and_ctes.events_table, (SELECT DISTINCT users_table.value_2 FROM subquery_and_ctes.users_table OFFSET 0) foo WHERE ((events_table.user_id OPERATOR(pg_catalog.=) foo.value_2) AND (events_table.user_id OPERATOR(pg_catalog.=) ANY (SELECT DISTINCT users_table.value_1 FROM subquery_and_ctes.users_table ORDER BY users_table.value_1 LIMIT 3)))) SELECT dist_cte.user_id FROM (local_cte JOIN dist_cte ON ((dist_cte.user_id OPERATOR(pg_catalog.=) local_cte.user_id)))
 DEBUG:  generating subplan 43_1 for CTE local_cte: SELECT user_id, "time", value_1, value_2, value_3, value_4 FROM subquery_and_ctes.users_table_local
@@ -440,41 +438,41 @@ DEBUG:  Plan 42 query after replacing subqueries and CTEs: SELECT count(*) AS co
 (1 row)
 
 -- the same query, but this time the CTEs also live inside a subquery
-SELECT 
-	* 
-FROM 
+SELECT
+	*
+FROM
 (
 	WITH cte AS (
 	WITH local_cte AS (
 		SELECT * FROM users_table_local
 	),
 	dist_cte AS (
-		SELECT 
+		SELECT
 			user_id
-		FROM 
-			events_table, 
+		FROM
+			events_table,
 			(SELECT DISTINCT value_2 FROM users_table OFFSET 0) as foo
-		WHERE 
+		WHERE
 			events_table.user_id = foo.value_2 AND
 			events_table.user_id IN (SELECT DISTINCT value_1 FROM users_table ORDER BY 1 LIMIT 3)
 	)
 	SELECT dist_cte.user_id FROM local_cte JOIN dist_cte ON dist_cte.user_id=local_cte.user_id
 )
-SELECT 
+SELECT
 	count(*)  as cnt
-FROM 
+FROM
 	cte,
-	  (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
+	  (SELECT
+    	DISTINCT users_table.user_id
+     FROM
      	users_table, events_table
-     WHERE 
-     	users_table.user_id = events_table.user_id AND 
+     WHERE
+     	users_table.user_id = events_table.user_id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
-     ) as foo 
+     ) as foo
 	  WHERE foo.user_id = cte.user_id
-) as foo, users_table WHERE foo.cnt > users_table.value_2 
+) as foo, users_table WHERE foo.cnt > users_table.value_2
 ORDER BY 3 DESC, 1 DESC, 2 DESC, 4 DESC
 LIMIT 5;
 DEBUG:  generating subplan 48_1 for CTE cte: WITH local_cte AS (SELECT users_table_local.user_id, users_table_local."time", users_table_local.value_1, users_table_local.value_2, users_table_local.value_3, users_table_local.value_4 FROM subquery_and_ctes.users_table_local), dist_cte AS (SELECT events_table.user_id FROM subquery_and_ctes.events_table, (SELECT DISTINCT users_table.value_2 FROM subquery_and_ctes.users_table OFFSET 0) foo WHERE ((events_table.user_id OPERATOR(pg_catalog.=) foo.value_2) AND (events_table.user_id OPERATOR(pg_catalog.=) ANY (SELECT DISTINCT users_table.value_1 FROM subquery_and_ctes.users_table ORDER BY users_table.value_1 LIMIT 3)))) SELECT dist_cte.user_id FROM (local_cte JOIN dist_cte ON ((dist_cte.user_id OPERATOR(pg_catalog.=) local_cte.user_id)))
@@ -505,25 +503,24 @@ SELECT
 FROM
     (
 	     WITH RECURSIVE cte AS (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
 	     ) SELECT * FROM cte ORDER BY 1 DESC
-     ) as foo, 
+     ) as foo,
     (
-	    SELECT 
-	    	DISTINCT users_table.user_id 
-	     FROM 
-	     	users_table, events_table 
-	     WHERE 
-	     	users_table.user_id = events_table.user_id AND 
+	    SELECT
+	    	DISTINCT users_table.user_id
+	     FROM
+	     	users_table, events_table
+	     WHERE
+	     	users_table.user_id = events_table.user_id AND
 	     event_type IN (1,2,3,4)
-	    
-     ) as bar  
+     ) as bar
 WHERE foo.user_id = bar.user_id
 ORDER BY 1 DESC;
 ERROR:  recursive CTEs are not supported in distributed queries

src/test/regress/expected/subquery_complex_target_list.out

@@ -1,5 +1,5 @@
 -- ===================================================================
--- test recursive planning functionality with complex target entries 
+-- test recursive planning functionality with complex target entries
 -- and some utilities
 -- ===================================================================
 CREATE SCHEMA subquery_complex;
@@ -11,7 +11,7 @@ SELECT
 FROM
   events_table
 WHERE
-  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20) 
+  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20)
 GROUP BY
   event_type
 ORDER BY 1 DESC, 2 DESC
@@ -63,8 +63,8 @@ FROM
     	SELECT count(distinct value_2) as cnt_2 FROM users_table ORDER BY 1 DESC LIMIT 4
     ) as baz,
 	(
-		SELECT user_id, sum(distinct value_2) as sum FROM users_table GROUP BY user_id ORDER BY 1 DESC LIMIT 4   
-    ) as bat, events_table 
+		SELECT user_id, sum(distinct value_2) as sum FROM users_table GROUP BY user_id ORDER BY 1 DESC LIMIT 4
+    ) as bat, events_table
     WHERE foo.avg != bar.cnt_1 AND baz.cnt_2 = events_table.event_type
     ORDER BY 1 DESC;
 DEBUG:  push down of limit count: 3
@@ -85,29 +85,29 @@ SELECT
    *
 FROM
     (
-    	SELECT 
+    	SELECT
     		min(user_id) * 2, max(user_id) / 2, sum(user_id), count(user_id)::float, avg(user_id)::bigint
-    	FROM 
-    		users_table 
-    	ORDER BY 1 DESC 
+    	FROM
+    		users_table
+    	ORDER BY 1 DESC
     	LIMIT 3
     ) as foo,
     (
-    	   SELECT 
+    	   SELECT
     		min(value_3) * 2, max(value_3) / 2, sum(value_3), count(value_3), avg(value_3)
-    	FROM 
-    		users_table 
-    	ORDER BY 1 DESC 
+    	FROM
+    		users_table
+    	ORDER BY 1 DESC
     	LIMIT 3
     ) as bar,
 	(
-    	SELECT 
-    		min(time), max(time), count(time), 
-    		count(*) FILTER (WHERE user_id = 3) as cnt_with_filter, 
+    	SELECT
+    		min(time), max(time), count(time),
+    		count(*) FILTER (WHERE user_id = 3) as cnt_with_filter,
     		count(*) FILTER (WHERE user_id::text LIKE '%3%') as cnt_with_filter_2
-    	FROM 
-    		users_table 
-    	ORDER BY 1 DESC 
+    	FROM
+    		users_table
+    	ORDER BY 1 DESC
     	LIMIT 3
     ) as baz
     ORDER BY 1 DESC;
@@ -135,11 +135,11 @@ FROM
     	SELECT sum(user_id * (5.0 / (value_1 + value_2 + 0.1)) * value_3) as cnt_1 FROM users_table ORDER BY 1 DESC LIMIT 3
     ) as bar,
 	(
-    	SELECT     
+    	SELECT
     		avg(case
             	when user_id > 4
             	then value_1
-        	end) as cnt_2, 
+        	end) as cnt_2,
     		avg(case
             	when user_id > 500
             	then value_1
@@ -152,16 +152,16 @@ FROM
 			end) as sum_1,
 			extract(year FROM max(time)) as l_year,
 			strpos(max(user_id)::text, '1') as pos
-         FROM 
-         		users_table 
-         ORDER BY 
-         	1 DESC 
+         FROM
+         		users_table
+         ORDER BY
+         	1 DESC
          LIMIT 4
-    ) as baz, 
+    ) as baz,
 	(
 		SELECT  COALESCE(value_3, 20) AS count_pay FROM users_table ORDER BY 1 OFFSET 20 LIMIT 5
 	) as tar,
-    events_table 
+    events_table
     WHERE foo.avg != bar.cnt_1 AND baz.cnt_2 != events_table.event_type
     ORDER BY 1 DESC;
 DEBUG:  push down of limit count: 3
@@ -245,7 +245,7 @@ FROM (
       		sum(value_1) > 10
       	   ORDER BY (sum(value_3) - avg(value_1) - COALESCE(array_upper(ARRAY[max(user_id)],1) * 5,0)) DESC
       	   LIMIT 3
-        ) as c 
+        ) as c
         WHERE b.value_2 != a.user_id
         ORDER BY 3 DESC, 2 DESC, 1 DESC
         LIMIT 5;
@@ -267,20 +267,20 @@ DEBUG:  Plan 21 query after replacing subqueries and CTEs: SELECT a.user_id, b.v
 SELECT
    bar.user_id
 FROM
-    (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
-     	users_table, events_table 
-     WHERE 
-     	users_table.user_id = events_table.user_id AND 
+    (SELECT
+    	DISTINCT users_table.user_id
+     FROM
+     	users_table, events_table
+     WHERE
+     	users_table.user_id = events_table.user_id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
      ) as foo,
-    (SELECT 
-    	DISTINCT users_table.user_id 
-     FROM 
-     	users_table, events_table 
-     WHERE 
+    (SELECT
+    	DISTINCT users_table.user_id
+     FROM
+     	users_table, events_table
+     WHERE
      	users_table.user_id = events_table.user_id AND false AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
@@ -296,7 +296,7 @@ DEBUG:  Plan 25 query after replacing subqueries and CTEs: SELECT bar.user_id FR
 (0 rows)
 
 -- window functions tests, both is recursively planned
-SELECT * FROM 
+SELECT * FROM
 (
 	SELECT
 	   user_id, time, rnk
@@ -325,7 +325,7 @@ SELECT * FROM
 	    *, rank() OVER my_win as rnk
 	  FROM
 	    events_table
-	  WHERE 
+	  WHERE
 	   	user_id = 3
 	  WINDOW my_win AS (PARTITION BY event_type ORDER BY time DESC)
 	) as foo
@@ -343,14 +343,13 @@ DEBUG:  Plan 28 query after replacing subqueries and CTEs: SELECT foo.user_id, f
 
 -- cursor test
 BEGIN;
-	
-	DECLARE recursive_subquery CURSOR FOR     
+	DECLARE recursive_subquery CURSOR FOR
 	SELECT
 	  event_type, count(distinct value_2)
 	FROM
 	  events_table
 	WHERE
-	  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20) 
+	  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20)
 	GROUP BY
 	  event_type
 	ORDER BY 1 DESC, 2 DESC
@@ -384,14 +383,13 @@ DEBUG:  Plan 31 query after replacing subqueries and CTEs: SELECT event_type, co
 COMMIT;
 -- cursor test with FETCH ALL
 BEGIN;
-	
-	DECLARE recursive_subquery CURSOR FOR     
+	DECLARE recursive_subquery CURSOR FOR
 	SELECT
 	  event_type, count(distinct value_2)
 	FROM
 	  events_table
 	WHERE
-	  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20) 
+	  user_id IN (SELECT user_id FROM users_table GROUP BY user_id ORDER BY count(*) DESC LIMIT 20)
 	GROUP BY
 	  event_type
 	ORDER BY 1 DESC, 2 DESC

src/test/regress/expected/subquery_in_where.out

@@ -4,13 +4,13 @@
 CREATE SCHEMA subquery_in_where;
 SET search_path TO subquery_in_where, public;
 SET client_min_messages TO DEBUG1;
---CTEs can be used as a recurring tuple with subqueries in WHERE 
+--CTEs can be used as a recurring tuple with subqueries in WHERE
 WITH event_id
-     AS (SELECT user_id AS events_user_id, 
-                time    AS events_time, 
+     AS (SELECT user_id AS events_user_id,
+                time    AS events_time,
                 event_type
-         FROM   events_table) 
-SELECT Count(*) 
+         FROM   events_table)
+SELECT Count(*)
 FROM   event_id
 WHERE  events_user_id IN (SELECT user_id
                           FROM   users_table);
@@ -22,27 +22,27 @@ DEBUG:  Plan 1 query after replacing subqueries and CTEs: SELECT count(*) AS cou
    101
 (1 row)
 
---Correlated subqueries can not be used in WHERE clause 
-WITH event_id 
-     AS (SELECT user_id AS events_user_id, 
-                time    AS events_time, 
-                event_type 
-         FROM   events_table) 
-SELECT Count(*) 
-FROM   event_id 
-WHERE  events_user_id IN (SELECT user_id 
-                          FROM   users_table 
-                          WHERE  users_table.time = events_time); 
+--Correlated subqueries can not be used in WHERE clause
+WITH event_id
+     AS (SELECT user_id AS events_user_id,
+                time    AS events_time,
+                event_type
+         FROM   events_table)
+SELECT Count(*)
+FROM   event_id
+WHERE  events_user_id IN (SELECT user_id
+                          FROM   users_table
+                          WHERE  users_table.time = events_time);
 DEBUG:  generating subplan 4_1 for CTE event_id: SELECT user_id AS events_user_id, "time" AS events_time, event_type FROM public.events_table
 DEBUG:  Plan 4 query after replacing subqueries and CTEs: SELECT count(*) AS count FROM (SELECT intermediate_result.events_user_id, intermediate_result.events_time, intermediate_result.event_type FROM read_intermediate_result('4_1'::text, 'binary'::citus_copy_format) intermediate_result(events_user_id integer, events_time timestamp without time zone, event_type integer)) event_id WHERE (events_user_id OPERATOR(pg_catalog.=) ANY (SELECT users_table.user_id FROM public.users_table WHERE (users_table."time" OPERATOR(pg_catalog.=) event_id.events_time)))
 ERROR:  cannot pushdown the subquery
 DETAIL:  Complex subqueries and CTEs are not allowed in the FROM clause when the query has subqueries in the WHERE clause and it references a column from another query
--- Recurring tuples as empty join tree 
-SELECT * 
+-- Recurring tuples as empty join tree
+SELECT *
 FROM   (SELECT 1 AS id, 2 AS value_1, 3 AS value_3
-		UNION ALL SELECT 2 as id, 3 as value_1, 4 as value_3) AS tt1 
-WHERE  id IN (SELECT user_id 
-              FROM   events_table); 
+		UNION ALL SELECT 2 as id, 3 as value_1, 4 as value_3) AS tt1
+WHERE  id IN (SELECT user_id
+              FROM   events_table);
 DEBUG:  generating subplan 6_1 for subquery SELECT 1 AS id, 2 AS value_1, 3 AS value_3 UNION ALL SELECT 2 AS id, 3 AS value_1, 4 AS value_3
 DEBUG:  generating subplan 6_2 for subquery SELECT user_id FROM public.events_table
 DEBUG:  Plan 6 query after replacing subqueries and CTEs: SELECT id, value_1, value_3 FROM (SELECT intermediate_result.id, intermediate_result.value_1, intermediate_result.value_3 FROM read_intermediate_result('6_1'::text, 'binary'::citus_copy_format) intermediate_result(id integer, value_1 integer, value_3 integer)) tt1 WHERE (id OPERATOR(pg_catalog.=) ANY (SELECT intermediate_result.user_id FROM read_intermediate_result('6_2'::text, 'binary'::citus_copy_format) intermediate_result(user_id integer)))
@@ -55,12 +55,12 @@ DEBUG:  Plan 6 query after replacing subqueries and CTEs: SELECT id, value_1, va
 -- Recurring tuples in from clause as CTE and SET operation in WHERE clause
 SELECT Count(*)
 FROM   (WITH event_id AS
-       (SELECT user_id AS events_user_id, time AS events_time, event_type 
+       (SELECT user_id AS events_user_id, time AS events_time, event_type
         FROM events_table)
        SELECT events_user_id, events_time, event_type
-	   FROM event_id 
+	   FROM event_id
 	   ORDER BY 1,2,3
-	   LIMIT 10) AS sub_table 
+	   LIMIT 10) AS sub_table
 WHERE  events_user_id IN (
        (SELECT user_id
         FROM users_table
@@ -218,7 +218,7 @@ FROM
 		user_id as events_user_id, time as events_time, event_type
 	FROM
 		events_table
-	ORDER BY 
+	ORDER BY
 		1,2,3
 	LIMIT
 		10
@@ -628,7 +628,6 @@ DEBUG:  Plan 65 query after replacing subqueries and CTEs: SELECT generate_serie
                3
 (3 rows)
 
-	
 -- Local tables also planned recursively, so using it as part of the FROM clause
 -- make the clause recurring
 CREATE TABLE local_table(id int, value_1 int);
@@ -656,7 +655,6 @@ DEBUG:  Plan 67 query after replacing subqueries and CTEs: SELECT id, value_1 FR
   2 |       2
 (2 rows)
 
-		
 -- Use local table in WHERE clause
 SELECT
 	COUNT(*)
@@ -667,7 +665,7 @@ FROM
 		users_table
 	ORDER BY
 		user_id
-	LIMIT 
+	LIMIT
 		10) as sub_table
 WHERE
 	user_id

src/test/regress/expected/subquery_partitioning.out

@@ -6,7 +6,6 @@ SET search_path TO subquery_and_partitioning, public;
 CREATE TABLE users_table_local AS SELECT * FROM users_table;
 CREATE TABLE events_table_local AS SELECT * FROM events_table;
 CREATE TABLE partitioning_test(id int, value_1 int, time date) PARTITION BY RANGE (time);
- 
 -- create its partitions
 CREATE TABLE partitioning_test_2017 PARTITION OF partitioning_test FOR VALUES FROM ('2017-01-01') TO ('2018-01-01');
 CREATE TABLE partitioning_test_2010 PARTITION OF partitioning_test FOR VALUES FROM ('2010-01-01') TO ('2011-01-01');
@@ -30,9 +29,9 @@ SET client_min_messages TO DEBUG1;
 SELECT
    id
 FROM
-    (SELECT 
-    	DISTINCT partitioning_test.id 
-     FROM 
+    (SELECT
+    	DISTINCT partitioning_test.id
+     FROM
      	partitioning_test
      LIMIT 5
      ) as foo
@@ -52,15 +51,15 @@ DEBUG:  Plan 3 query after replacing subqueries and CTEs: SELECT id FROM (SELECT
 SELECT
    *
 FROM
-    (SELECT 
-    	DISTINCT partitioning_test.id 
-     FROM 
+    (SELECT
+    	DISTINCT partitioning_test.id
+     FROM
      	partitioning_test
      LIMIT 5
      ) as foo,
-	(SELECT 
-    	DISTINCT partitioning_test.time 
-     FROM 
+	(SELECT
+    	DISTINCT partitioning_test.time
+     FROM
      	partitioning_test
      LIMIT 5
      ) as bar
@@ -80,17 +79,17 @@ DEBUG:  Plan 5 query after replacing subqueries and CTEs: SELECT foo.id, bar."ti
 SELECT
    *
 FROM
-    (SELECT 
-    	DISTINCT partitioning_test.time 
-     FROM 
+    (SELECT
+    	DISTINCT partitioning_test.time
+     FROM
      	partitioning_test
  	 ORDER BY 1 DESC
      LIMIT 5
      ) as foo,
 	(
-		SELECT 
-	    	DISTINCT partitioning_test.id 
-	     FROM 
+		SELECT
+	    	DISTINCT partitioning_test.id
+	     FROM
 	     	partitioning_test
      ) as bar
 	WHERE  date_part('day', foo.time) = bar.id
@@ -109,19 +108,19 @@ DEBUG:  push down of limit count: 3
 SELECT
    *
 FROM
-    (SELECT 
-    	DISTINCT partitioning_test.time 
-     FROM 
+    (SELECT
+    	DISTINCT partitioning_test.time
+     FROM
      	partitioning_test
  	 ORDER BY 1 DESC
      LIMIT 5
      ) as foo,
 	(
-		SELECT 
-	    	DISTINCT partitioning_test.id 
-	     FROM 
+		SELECT
+	    	DISTINCT partitioning_test.id
+	     FROM
 	     	partitioning_test
-     ) as bar, 
+     ) as bar,
 	partitioning_test
 	WHERE  date_part('day', foo.time) = bar.id AND partitioning_test.id = bar.id
 	ORDER BY 2 DESC, 1 DESC
@@ -138,7 +137,7 @@ DEBUG:  push down of limit count: 3
 -- subquery in WHERE clause
 SELECT DISTINCT id
 FROM partitioning_test
-WHERE 
+WHERE
 	id IN (SELECT DISTINCT date_part('day', time) FROM partitioning_test);
 DEBUG:  generating subplan 12_1 for subquery SELECT DISTINCT date_part('day'::text, "time") AS date_part FROM subquery_and_partitioning.partitioning_test
 DEBUG:  Plan 12 query after replacing subqueries and CTEs: SELECT DISTINCT id FROM subquery_and_partitioning.partitioning_test WHERE ((id)::double precision OPERATOR(pg_catalog.=) ANY (SELECT intermediate_result.date_part FROM read_intermediate_result('12_1'::text, 'binary'::citus_copy_format) intermediate_result(date_part double precision)))
@@ -149,16 +148,16 @@ DEBUG:  Plan 12 query after replacing subqueries and CTEs: SELECT DISTINCT id FR
 
 -- repartition subquery
 SET citus.enable_repartition_joins to ON;
-SELECT 
-	count(*) 
+SELECT
+	count(*)
 FROM
 (
 	SELECT DISTINCT p1.value_1 FROM partitioning_test as p1, partitioning_test as p2 WHERE p1.id = p2.value_1
-) as foo, 
+) as foo,
 (
 	SELECT user_id FROM users_table
 ) as bar
-WHERE foo.value_1 = bar.user_id; 
+WHERE foo.value_1 = bar.user_id;
 DEBUG:  cannot use adaptive executor with repartition jobs
 HINT:  Since you enabled citus.enable_repartition_joins Citus chose to use task-tracker.
 DEBUG:  generating subplan 14_1 for subquery SELECT DISTINCT p1.value_1 FROM subquery_and_partitioning.partitioning_test p1, subquery_and_partitioning.partitioning_test p2 WHERE (p1.id OPERATOR(pg_catalog.=) p2.value_1)
@@ -170,40 +169,40 @@ DEBUG:  Plan 14 query after replacing subqueries and CTEs: SELECT count(*) AS co
 
 SET citus.enable_repartition_joins to OFF;
 -- subquery, cte, view and non-partitioned tables
-CREATE VIEW subquery_and_ctes AS 
-SELECT 
-	* 
-FROM 
+CREATE VIEW subquery_and_ctes AS
+SELECT
+	*
+FROM
 (
 	WITH cte AS (
 	WITH local_cte AS (
 		SELECT * FROM users_table_local
 	),
 	dist_cte AS (
-		SELECT 
+		SELECT
 			user_id
-		FROM 
-			events_table, 
+		FROM
+			events_table,
 			(SELECT DISTINCT value_1 FROM partitioning_test OFFSET 0) as foo
-		WHERE 
+		WHERE
 			events_table.user_id = foo.value_1 AND
 			events_table.user_id IN (SELECT DISTINCT value_1 FROM users_table ORDER BY 1 LIMIT 3)
 	)
 	SELECT dist_cte.user_id FROM local_cte join dist_cte on dist_cte.user_id=local_cte.user_id
 )
-SELECT 
+SELECT
 	count(*)  as cnt
-FROM 
+FROM
 	cte,
-	  (SELECT 
-    	DISTINCT events_table.user_id 
-     FROM 
+	  (SELECT
+    	DISTINCT events_table.user_id
+     FROM
      	partitioning_test, events_table
-     WHERE 
-     	events_table.user_id = partitioning_test.id AND 
+     WHERE
+     	events_table.user_id = partitioning_test.id AND
      event_type IN (1,2,3,4)
      ORDER BY 1 DESC LIMIT 5
-     ) as foo 
+     ) as foo
 	  WHERE foo.user_id = cte.user_id
 ) as foo, users_table WHERE foo.cnt > users_table.value_2;
 SELECT * FROM subquery_and_ctes
@@ -235,27 +234,27 @@ DEBUG:  push down of limit count: 5
 SELECT count(*)
 FROM
 (
-	SELECT avg(min) FROM 
+	SELECT avg(min) FROM
 	(
 		SELECT min(partitioning_test.value_1) FROM
 		(
-			SELECT avg(event_type) as avg_ev_type FROM 
+			SELECT avg(event_type) as avg_ev_type FROM
 			(
-				SELECT 
-					max(value_1) as mx_val_1 
+				SELECT
+					max(value_1) as mx_val_1
 					FROM (
-							SELECT 
+							SELECT
 								avg(event_type) as avg
 							FROM
 							(
-								SELECT 
-									cnt 
-								FROM 
+								SELECT
+									cnt
+								FROM
 									(SELECT count(*) as cnt, value_1 FROM partitioning_test GROUP BY value_1) as level_1, users_table
-								WHERE 
+								WHERE
 									users_table.user_id = level_1.cnt
 							) as level_2, events_table
-							WHERE events_table.user_id = level_2.cnt 
+							WHERE events_table.user_id = level_2.cnt
 							GROUP BY level_2.cnt
 						) as level_3, users_table
 					WHERE user_id = level_3.avg
@@ -264,9 +263,9 @@ FROM
 				WHERE level_4.mx_val_1 = events_table.user_id
 				GROUP BY level_4.mx_val_1
 				) as level_5, partitioning_test
-				WHERE 
+				WHERE
 					level_5.avg_ev_type = partitioning_test.id
-				GROUP BY 
+				GROUP BY
 					level_5.avg_ev_type
 		) as level_6, users_table WHERE users_table.user_id = level_6.min
 	GROUP BY users_table.value_1

src/test/regress/expected/with_dml.out

@@ -29,7 +29,7 @@ SET client_min_messages TO DEBUG1;
 WITH ids_to_delete AS (
   SELECT tenant_id FROM distributed_table WHERE dept = 1
 )
-DELETE FROM reference_table WHERE id IN (SELECT tenant_id FROM ids_to_delete); 
+DELETE FROM reference_table WHERE id IN (SELECT tenant_id FROM ids_to_delete);
 DEBUG:  generating subplan 4_1 for CTE ids_to_delete: SELECT tenant_id FROM with_dml.distributed_table WHERE (dept OPERATOR(pg_catalog.=) 1)
 DEBUG:  Plan 4 query after replacing subqueries and CTEs: DELETE FROM with_dml.reference_table WHERE (id OPERATOR(pg_catalog.=) ANY (SELECT ids_to_delete.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('4_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_to_delete))
 -- update the name of the users whose dept is 2
@@ -40,11 +40,11 @@ UPDATE reference_table SET name = 'new_' || name WHERE id IN (SELECT tenant_id F
 DEBUG:  generating subplan 6_1 for CTE ids_to_update: SELECT tenant_id FROM with_dml.distributed_table WHERE (dept OPERATOR(pg_catalog.=) 2)
 DEBUG:  Plan 6 query after replacing subqueries and CTEs: UPDATE with_dml.reference_table SET name = ('new_'::text OPERATOR(pg_catalog.||) name) WHERE (id OPERATOR(pg_catalog.=) ANY (SELECT ids_to_update.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('6_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_to_update))
 -- now the CTE is also modifying
-WITH ids_deleted_3 AS 
+WITH ids_deleted_3 AS
 (
 	DELETE FROM distributed_table WHERE dept = 3 RETURNING tenant_id
 ),
-ids_deleted_4 AS 
+ids_deleted_4 AS
 (
 	DELETE FROM distributed_table WHERE dept = 4 RETURNING tenant_id
 )
@@ -54,15 +54,15 @@ DEBUG:  generating subplan 8_2 for CTE ids_deleted_4: DELETE FROM with_dml.distr
 DEBUG:  generating subplan 8_3 for subquery SELECT ids_deleted_3.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('8_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_deleted_3 UNION SELECT ids_deleted_4.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('8_2'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_deleted_4
 DEBUG:  Plan 8 query after replacing subqueries and CTEs: DELETE FROM with_dml.reference_table WHERE (id OPERATOR(pg_catalog.=) ANY (SELECT intermediate_result.tenant_id FROM read_intermediate_result('8_3'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)))
 -- now the final UPDATE command is pushdownable
-WITH ids_to_delete AS 
+WITH ids_to_delete AS
 (
 	SELECT tenant_id FROM distributed_table WHERE dept = 5
 )
-UPDATE 
-	distributed_table 
-SET 
+UPDATE
+	distributed_table
+SET
 	dept = dept + 1
-FROM 
+FROM
 	ids_to_delete, (SELECT tenant_id FROM distributed_table WHERE tenant_id::int < 60) as some_tenants
 WHERE
 	some_tenants.tenant_id = ids_to_delete.tenant_id
@@ -72,25 +72,25 @@ DEBUG:  generating subplan 12_1 for CTE ids_to_delete: SELECT tenant_id FROM wit
 DEBUG:  Plan 12 query after replacing subqueries and CTEs: UPDATE with_dml.distributed_table SET dept = (distributed_table.dept OPERATOR(pg_catalog.+) 1) FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('12_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_to_delete, (SELECT distributed_table_1.tenant_id FROM with_dml.distributed_table distributed_table_1 WHERE ((distributed_table_1.tenant_id)::integer OPERATOR(pg_catalog.<) 60)) some_tenants WHERE ((some_tenants.tenant_id OPERATOR(pg_catalog.=) ids_to_delete.tenant_id) AND (distributed_table.tenant_id OPERATOR(pg_catalog.=) some_tenants.tenant_id) AND (EXISTS (SELECT ids_to_delete_1.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('12_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text)) ids_to_delete_1)))
 -- this query errors out since we've some hard
 -- errors in the INSERT ... SELECT pushdown
--- which prevents to fallback to recursive planning 
-WITH ids_to_upsert AS 
+-- which prevents to fallback to recursive planning
+WITH ids_to_upsert AS
 (
 	SELECT tenant_id FROM distributed_table WHERE dept > 7
 )
-INSERT INTO distributed_table  
+INSERT INTO distributed_table
        SELECT distributed_table.tenant_id FROM ids_to_upsert, distributed_table
        		WHERE  distributed_table.tenant_id = ids_to_upsert.tenant_id
        	ON CONFLICT (tenant_id) DO UPDATE SET dept = 8;
 ERROR:  cannot perform distributed planning for the given modification
 DETAIL:  Select query cannot be pushed down to the worker.
 -- the following query is very similar to the above one
--- but this time the query is pulled to coordinator since 
+-- but this time the query is pulled to coordinator since
 -- we return before hitting any hard errors
-WITH ids_to_insert AS 
+WITH ids_to_insert AS
 (
 	SELECT (tenant_id::int * 100)::text as tenant_id FROM distributed_table WHERE dept > 7
 )
-INSERT INTO distributed_table  
+INSERT INTO distributed_table
        SELECT DISTINCT ids_to_insert.tenant_id FROM ids_to_insert, distributed_table
        		WHERE  distributed_table.tenant_id < ids_to_insert.tenant_id;
 DEBUG:  cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
@@ -108,7 +108,7 @@ DEBUG:  Plan 16 query after replacing subqueries and CTEs: SELECT tenant_id FROM
 -- since COPY cannot be executed
 SET citus.force_max_query_parallelization TO on;
 WITH copy_to_other_table AS (
-    INSERT INTO distributed_table 
+    INSERT INTO distributed_table
         SELECT *
             FROM second_distributed_table
         WHERE dept = 3
@@ -116,10 +116,10 @@ WITH copy_to_other_table AS (
         RETURNING *
 ),
 main_table_deleted AS (
-    DELETE 
-    FROM distributed_table 
+    DELETE
+    FROM distributed_table
     WHERE dept < 10
-      AND NOT EXISTS (SELECT 1 FROM second_distributed_table 
+      AND NOT EXISTS (SELECT 1 FROM second_distributed_table
                       WHERE second_distributed_table.dept = 1
                         AND second_distributed_table.tenant_id = distributed_table.tenant_id)
                         RETURNING *
@@ -127,7 +127,7 @@ main_table_deleted AS (
 INSERT INTO second_distributed_table
         SELECT *
             FROM main_table_deleted
-        EXCEPT 
+        EXCEPT
         SELECT *
             FROM copy_to_other_table;
 DEBUG:  distributed INSERT ... SELECT can only select from distributed tables
@@ -138,20 +138,19 @@ DEBUG:  generating subplan 20_3 for subquery SELECT main_table_deleted.tenant_id
 DEBUG:  Plan 20 query after replacing subqueries and CTEs: SELECT tenant_id, dept FROM (SELECT intermediate_result.tenant_id, intermediate_result.dept FROM read_intermediate_result('20_3'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id text, dept integer)) citus_insert_select_subquery
 SET citus.force_max_query_parallelization TO off;
 -- CTE inside the UPDATE statement
-UPDATE 
-	second_distributed_table 
-SET dept = 
+UPDATE
+	second_distributed_table
+SET dept =
 		(WITH  vals AS (
 		SELECT DISTINCT tenant_id::int FROM distributed_table
-		) select * from vals where tenant_id = 8 ) 
+		) select * from vals where tenant_id = 8 )
 		WHERE dept = 8;
 DEBUG:  generating subplan 24_1 for CTE vals: SELECT DISTINCT (tenant_id)::integer AS tenant_id FROM with_dml.distributed_table
 DEBUG:  Plan 24 query after replacing subqueries and CTEs: UPDATE with_dml.second_distributed_table SET dept = (SELECT vals.tenant_id FROM (SELECT intermediate_result.tenant_id FROM read_intermediate_result('24_1'::text, 'binary'::citus_copy_format) intermediate_result(tenant_id integer)) vals WHERE (vals.tenant_id OPERATOR(pg_catalog.=) 8)) WHERE (dept OPERATOR(pg_catalog.=) 8)
 -- Subquery inside the UPDATE statement
-UPDATE 
-	second_distributed_table 
-SET dept = 
-		
+UPDATE
+	second_distributed_table
+SET dept =
 		(SELECT DISTINCT tenant_id::int FROM distributed_table WHERE tenant_id = '9')
 		WHERE dept = 8;
 DEBUG:  generating subplan 26_1 for subquery SELECT DISTINCT (tenant_id)::integer AS tenant_id FROM with_dml.distributed_table WHERE (tenant_id OPERATOR(pg_catalog.=) '9'::text)

src/test/regress/expected/with_join.out

@@ -10,9 +10,9 @@ SELECT create_reference_table('with_join.reference_table');
 
 INSERT INTO reference_table VALUES (6), (7);
 SET citus.enable_repartition_joins TO on;
--- Two colocated CTE under a non-colocated join 
+-- Two colocated CTE under a non-colocated join
 WITH colocated_1 AS (
-  SELECT 
+  SELECT
     users_table.user_id, events_table.value_2
   FROM
     users_table, events_table
@@ -20,14 +20,14 @@ WITH colocated_1 AS (
     users_table.user_id = events_table.user_id AND event_type IN (1, 2, 3)
 ),
 colocated_2 AS (
-  SELECT 
+  SELECT
     users_table.user_id, events_table.value_2
   FROM
     users_table, events_table
   WHERE
     users_table.user_id = events_table.user_id AND event_type IN (4, 5, 6)
 )
-SELECT colocated_1.user_id, count(*) 
+SELECT colocated_1.user_id, count(*)
 FROM
   colocated_1, colocated_2
 WHERE
@@ -46,9 +46,9 @@ ORDER BY
        6 |  6710
 (6 rows)
 
--- Two non-colocated CTE under a co-located join 
+-- Two non-colocated CTE under a co-located join
 WITH non_colocated_1 AS (
-  SELECT 
+  SELECT
     users_table.user_id
   FROM
     users_table, events_table
@@ -56,14 +56,14 @@ WITH non_colocated_1 AS (
     users_table.user_id = events_table.value_2 AND event_type IN (1, 2, 3)
 ),
 non_colocated_2 AS (
-  SELECT 
+  SELECT
     users_table.user_id
   FROM
     users_table, events_table
   WHERE
     users_table.user_id = events_table.value_2 AND event_type IN (4, 5, 6)
 )
-SELECT non_colocated_1.user_id, count(*) 
+SELECT non_colocated_1.user_id, count(*)
 FROM
   non_colocated_1, non_colocated_2
 WHERE
@@ -81,7 +81,6 @@ ORDER BY
        1 |  6272
 (5 rows)
 
-  
 -- Subqueries in WHERE and FROM are mixed
 -- In this query, only subquery in WHERE is not a colocated join
 -- but we're able to recursively plan that as well
@@ -121,12 +120,12 @@ WITH users_events AS (
         event_type
       FROM
         events_table
-      WHERE 
-        user_id < 100 
-      GROUP BY 
-        1 
-      ORDER BY 
-        1 
+      WHERE
+        user_id < 100
+      GROUP BY
+        1
+      ORDER BY
+        1
       LIMIT 10
     )
     SELECT
@@ -139,9 +138,9 @@ SELECT
   DISTINCT uid
 FROM
   users_events
-ORDER BY 
+ORDER BY
   1 DESC
-LIMIT 
+LIMIT
   5;
  uid 
 -----
@@ -162,9 +161,9 @@ FROM
   cte
 LEFT JOIN
   events_table ON cte.user_id = events_table.user_id
-ORDER BY 
+ORDER BY
   1,2,3
-LIMIT 
+LIMIT
   5;
 ERROR:  cannot pushdown the subquery
 DETAIL:  Complex subqueries and CTEs cannot be in the outer part of the outer join
@@ -178,9 +177,9 @@ FROM
   cte
 RIGHT JOIN
   events_table ON cte.user_id = events_table.user_id
-ORDER BY 
+ORDER BY
   1,2,3
-LIMIT 
+LIMIT
   5;
  user_id |              time               | event_type 
 ---------+---------------------------------+------------
@@ -198,12 +197,12 @@ WITH cte AS (
 SELECT
   cte.user_id, cte.time, events_table.event_type
 FROM
-  events_table 
+  events_table
 LEFT JOIN
   cte ON cte.user_id = events_table.user_id
-ORDER BY 
+ORDER BY
   1,2,3
-LIMIT 
+LIMIT
   5;
  user_id |              time               | event_type 
 ---------+---------------------------------+------------
@@ -221,12 +220,12 @@ WITH cte AS (
 SELECT
   cte.user_id, cte.time, events_table.event_type
 FROM
-  events_table 
+  events_table
 RIGHT JOIN
   cte ON cte.user_id = events_table.user_id
-ORDER BY 
+ORDER BY
   1,2,3
-LIMIT 
+LIMIT
   5;
 ERROR:  cannot pushdown the subquery
 DETAIL:  Complex subqueries and CTEs cannot be in the outer part of the outer join
@@ -237,12 +236,12 @@ WITH cte AS (
 SELECT
   cte.user_id, cte.time, events_table.event_type
 FROM
-  events_table 
+  events_table
 FULL JOIN
   cte ON cte.user_id = events_table.user_id
-ORDER BY 
+ORDER BY
   1,2,3
-LIMIT 
+LIMIT
   5;
 ERROR:  cannot pushdown the subquery
 DETAIL:  Complex subqueries and CTEs cannot be in the outer part of the outer join

src/test/regress/expected/with_nested.out

@@ -42,9 +42,9 @@ WITH users_events AS (
   WHERE
     u_events.user_id = events_table.user_id
   )
-SELECT 
-  * 
-FROM 
+SELECT
+  *
+FROM
   users_events
 ORDER BY
   1, 2
@@ -92,17 +92,16 @@ WITH users_events AS (
               GROUP BY
                   users_table.user_id,
                   events_table.event_type
-            
             )
-            SELECT 
+            SELECT
                 uid, event_type, value_2, value_3
             FROM
                (
                 (
-                  SELECT 
+                  SELECT
                     user_id as uid
-                  FROM 
-                    users_events 
+                  FROM
+                    users_events
                 ) users
                 join
                   events_table
@@ -110,45 +109,45 @@ WITH users_events AS (
                   users.uid = events_table.event_type
                 ) a
             )
-          SELECT 
-            * 
-          FROM 
+          SELECT
+            *
+          FROM
             users_events
           ORDER BY
             1, 3, 2, 4
           LIMIT 100
         )
-        SELECT 
-          * 
-        FROM 
+        SELECT
+          *
+        FROM
           users_events
         LIMIT 90
       )
-      SELECT 
-        * 
-      FROM 
+      SELECT
+        *
+      FROM
         users_events
       LIMIT 50
     )
-    SELECT 
+    SELECT
       uid, event_type, value_2, sum(value_3) as sum_3
-    FROM 
+    FROM
       users_events
     GROUP BY
       1, 2, 3
     LIMIT 40
   )
-  SELECT 
+  SELECT
     uid, event_type, sum(value_2) as sum_2, sum(sum_3) as sum_3
-  FROM 
+  FROM
     users_events
   GROUP BY
     1, 2
-  LIMIT 30   
+  LIMIT 30
 )
-SELECT 
+SELECT
   uid, avg(event_type), sum(sum_2), sum(sum_3)
-FROM 
+FROM
   users_events
 GROUP BY
   1;
@@ -210,7 +209,7 @@ WITH users_events AS (
         *
       FROM
         users_events_2_3
-    UNION 
+    UNION
       SELECT
         *
       FROM
@@ -221,8 +220,8 @@ WITH users_events AS (
   FROM
     merged_users
 )
-SELECT 
-  * 
+SELECT
+  *
 FROM
   users_events
 ORDER BY
@@ -305,14 +304,14 @@ WITH users_events AS (
         *
       FROM
         users_events_2_3
-  UNION 
+  UNION
       SELECT
         *
      FROM
        users_events_4
 )
-SELECT 
-  * 
+SELECT
+  *
 FROM
   users_events
 ORDER BY