-- -- MULTI_EXPLAIN -- ALTER SEQUENCE pg_catalog.pg_dist_shardid_seq RESTART 570000; ALTER SEQUENCE pg_catalog.pg_dist_jobid_seq RESTART 570000; \a\t SET citus.task_executor_type TO 'real-time'; SET citus.explain_distributed_queries TO on; -- Function that parses explain output as JSON CREATE FUNCTION explain_json(query text) RETURNS jsonb AS $BODY$ DECLARE result jsonb; BEGIN EXECUTE format('EXPLAIN (FORMAT JSON) %s', query) INTO result; RETURN result; END; $BODY$ LANGUAGE plpgsql; -- Function that parses explain output as XML CREATE FUNCTION explain_xml(query text) RETURNS xml AS $BODY$ DECLARE result xml; BEGIN EXECUTE format('EXPLAIN (FORMAT XML) %s', query) INTO result; RETURN result; END; $BODY$ LANGUAGE plpgsql; -- Test Text format EXPLAIN (COSTS FALSE, FORMAT TEXT) SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity; Distributed Query into pg_merge_job_570000 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> HashAggregate Group Key: l_quantity -> Seq Scan on lineitem_290001 lineitem Master Query -> Sort Sort Key: COALESCE((sum((COALESCE((sum(intermediate_column_570000_1))::bigint, '0'::bigint))))::bigint, '0'::bigint), intermediate_column_570000_0 -> HashAggregate Group Key: intermediate_column_570000_0 -> Seq Scan on pg_merge_job_570000 -- Test JSON format EXPLAIN (COSTS FALSE, FORMAT JSON) SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity; [ { "Executor": "Real-Time", "Job": { "Task Count": 8, "Tasks Shown": "One of 8", "Tasks": [ { "Node": "host=localhost port=57637 dbname=regression", "Remote Plan": [ [ { "Plan": { "Node Type": "Aggregate", "Strategy": "Hashed", "Group Key": ["l_quantity"], "Plans": [ { "Node Type": "Seq Scan", "Parent Relationship": "Outer", "Relation Name": "lineitem_290001", "Alias": "lineitem" } ] } } ] ] } ] }, "Master Query": [ { "Plan": { "Node Type": "Sort", "Sort Key": ["COALESCE((sum((COALESCE((sum(intermediate_column_570001_1))::bigint, '0'::bigint))))::bigint, '0'::bigint)", "intermediate_column_570001_0"], "Plans": [ { "Node Type": "Aggregate", "Strategy": "Hashed", "Parent Relationship": "Outer", "Group Key": ["intermediate_column_570001_0"], "Plans": [ { "Node Type": "Seq Scan", "Parent Relationship": "Outer", "Relation Name": "pg_merge_job_570001", "Alias": "pg_merge_job_570001" } ] } ] } } ] } ] -- Validate JSON format SELECT true AS valid FROM explain_json($$ SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity$$); t -- Test XML format EXPLAIN (COSTS FALSE, FORMAT XML) SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity; Real-Time 8 One of 8 host=localhost port=57637 dbname=regression Aggregate Hashed l_quantity Seq Scan Outer lineitem_290001 lineitem Sort COALESCE((sum((COALESCE((sum(intermediate_column_570003_1))::bigint, '0'::bigint))))::bigint, '0'::bigint) intermediate_column_570003_0 Aggregate Hashed Outer intermediate_column_570003_0 Seq Scan Outer pg_merge_job_570003 pg_merge_job_570003 -- Validate XML format SELECT true AS valid FROM explain_xml($$ SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity$$); t -- Test YAML format EXPLAIN (COSTS FALSE, FORMAT YAML) SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity; - Executor: "Real-Time" Job: Task Count: 8 Tasks Shown: "One of 8" Tasks: - Node: "host=localhost port=57637 dbname=regression" Remote Plan: - Plan: Node Type: "Aggregate" Strategy: "Hashed" Group Key: - "l_quantity" Plans: - Node Type: "Seq Scan" Parent Relationship: "Outer" Relation Name: "lineitem_290001" Alias: "lineitem" Master Query: - Plan: Node Type: "Sort" Sort Key: - "COALESCE((sum((COALESCE((sum(intermediate_column_570005_1))::bigint, '0'::bigint))))::bigint, '0'::bigint)" - "intermediate_column_570005_0" Plans: - Node Type: "Aggregate" Strategy: "Hashed" Parent Relationship: "Outer" Group Key: - "intermediate_column_570005_0" Plans: - Node Type: "Seq Scan" Parent Relationship: "Outer" Relation Name: "pg_merge_job_570005" Alias: "pg_merge_job_570005" -- Test Text format EXPLAIN (COSTS FALSE, FORMAT TEXT) SELECT l_quantity, count(*) count_quantity FROM lineitem GROUP BY l_quantity ORDER BY count_quantity, l_quantity; Distributed Query into pg_merge_job_570006 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> HashAggregate Group Key: l_quantity -> Seq Scan on lineitem_290001 lineitem Master Query -> Sort Sort Key: COALESCE((sum((COALESCE((sum(intermediate_column_570006_1))::bigint, '0'::bigint))))::bigint, '0'::bigint), intermediate_column_570006_0 -> HashAggregate Group Key: intermediate_column_570006_0 -> Seq Scan on pg_merge_job_570006 -- Test verbose EXPLAIN (COSTS FALSE, VERBOSE TRUE) SELECT sum(l_quantity) / avg(l_quantity) FROM lineitem; Distributed Query into pg_merge_job_570007 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> Aggregate Output: sum(l_quantity), sum(l_quantity), count(l_quantity) -> Seq Scan on public.lineitem_290001 lineitem Output: l_orderkey, l_partkey, l_suppkey, l_linenumber, l_quantity, l_extendedprice, l_discount, l_tax, l_returnflag, l_linestatus, l_shipdate, l_commitdate, l_receiptdate, l_shipinstruct, l_shipmode, l_comment Master Query -> Aggregate Output: (sum(intermediate_column_570007_0) / (sum(intermediate_column_570007_1) / sum(intermediate_column_570007_2))) -> Seq Scan on pg_temp_2.pg_merge_job_570007 Output: intermediate_column_570007_0, intermediate_column_570007_1, intermediate_column_570007_2 -- Test join EXPLAIN (COSTS FALSE) SELECT * FROM lineitem JOIN orders ON l_orderkey = o_orderkey AND l_quantity < 5.0 ORDER BY l_quantity LIMIT 10; Distributed Query into pg_merge_job_570008 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> Limit -> Sort Sort Key: lineitem.l_quantity -> Hash Join Hash Cond: (lineitem.l_orderkey = orders.o_orderkey) -> Seq Scan on lineitem_290001 lineitem Filter: (l_quantity < 5.0) -> Hash -> Seq Scan on orders_290008 orders Master Query -> Limit -> Sort Sort Key: intermediate_column_570008_4 -> Seq Scan on pg_merge_job_570008 -- Test insert EXPLAIN (COSTS FALSE) INSERT INTO lineitem VALUES(1,0); Distributed Query Executor: Router Task Count: 1 Tasks Shown: All -> Task Node: host=localhost port=57638 dbname=regression -> Insert on lineitem_290000 -> Result -- Test update EXPLAIN (COSTS FALSE) UPDATE lineitem SET l_suppkey = 12 WHERE l_orderkey = 1 AND l_partkey = 0; Distributed Query Executor: Router Task Count: 1 Tasks Shown: All -> Task Node: host=localhost port=57638 dbname=regression -> Update on lineitem_290000 -> Bitmap Heap Scan on lineitem_290000 Recheck Cond: (l_orderkey = 1) Filter: (l_partkey = 0) -> Bitmap Index Scan on lineitem_pkey_290000 Index Cond: (l_orderkey = 1) -- Test delete EXPLAIN (COSTS FALSE) DELETE FROM lineitem WHERE l_orderkey = 1 AND l_partkey = 0; Distributed Query Executor: Router Task Count: 1 Tasks Shown: All -> Task Node: host=localhost port=57638 dbname=regression -> Delete on lineitem_290000 -> Bitmap Heap Scan on lineitem_290000 Recheck Cond: (l_orderkey = 1) Filter: (l_partkey = 0) -> Bitmap Index Scan on lineitem_pkey_290000 Index Cond: (l_orderkey = 1) -- Test single-shard SELECT EXPLAIN (COSTS FALSE) SELECT l_quantity FROM lineitem WHERE l_orderkey = 5; Distributed Query into pg_merge_job_570009 Executor: Router Task Count: 1 Tasks Shown: All -> Task Node: host=localhost port=57637 dbname=regression -> Bitmap Heap Scan on lineitem_290000 lineitem Recheck Cond: (l_orderkey = 5) -> Bitmap Index Scan on lineitem_pkey_290000 Index Cond: (l_orderkey = 5) SELECT true AS valid FROM explain_xml($$ SELECT l_quantity FROM lineitem WHERE l_orderkey = 5$$); t SELECT true AS valid FROM explain_json($$ SELECT l_quantity FROM lineitem WHERE l_orderkey = 5$$); t -- Test CREATE TABLE ... AS EXPLAIN (COSTS FALSE) CREATE TABLE explain_result AS SELECT * FROM lineitem; Distributed Query into pg_merge_job_570012 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> Seq Scan on lineitem_290001 lineitem Master Query -> Seq Scan on pg_merge_job_570012 -- Test having EXPLAIN (COSTS FALSE, VERBOSE TRUE) SELECT sum(l_quantity) / avg(l_quantity) FROM lineitem HAVING sum(l_quantity) > 100; Distributed Query into pg_merge_job_570013 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> Aggregate Output: sum(l_quantity), sum(l_quantity), count(l_quantity), sum(l_quantity) -> Seq Scan on public.lineitem_290001 lineitem Output: l_orderkey, l_partkey, l_suppkey, l_linenumber, l_quantity, l_extendedprice, l_discount, l_tax, l_returnflag, l_linestatus, l_shipdate, l_commitdate, l_receiptdate, l_shipinstruct, l_shipmode, l_comment Master Query -> Aggregate Output: (sum(intermediate_column_570013_0) / (sum(intermediate_column_570013_1) / sum(intermediate_column_570013_2))) Filter: (sum(pg_merge_job_570013.intermediate_column_570013_3) > '100'::numeric) -> Seq Scan on pg_temp_2.pg_merge_job_570013 Output: intermediate_column_570013_0, intermediate_column_570013_1, intermediate_column_570013_2, intermediate_column_570013_3 -- Test having without aggregate EXPLAIN (COSTS FALSE, VERBOSE TRUE) SELECT l_quantity FROM lineitem GROUP BY l_quantity HAVING l_quantity > (100 * random()); Distributed Query into pg_merge_job_570014 Executor: Real-Time Task Count: 8 Tasks Shown: One of 8 -> Task Node: host=localhost port=57637 dbname=regression -> HashAggregate Output: l_quantity, l_quantity Group Key: lineitem.l_quantity -> Seq Scan on public.lineitem_290001 lineitem Output: l_orderkey, l_partkey, l_suppkey, l_linenumber, l_quantity, l_extendedprice, l_discount, l_tax, l_returnflag, l_linestatus, l_shipdate, l_commitdate, l_receiptdate, l_shipinstruct, l_shipmode, l_comment Master Query -> HashAggregate Output: intermediate_column_570014_0 Group Key: pg_merge_job_570014.intermediate_column_570014_0 Filter: ((pg_merge_job_570014.intermediate_column_570014_1)::double precision > ('100'::double precision * random())) -> Seq Scan on pg_temp_2.pg_merge_job_570014 Output: intermediate_column_570014_0, intermediate_column_570014_1 -- Test all tasks output SET citus.explain_all_tasks TO on; EXPLAIN (COSTS FALSE) SELECT avg(l_linenumber) FROM lineitem WHERE l_orderkey > 9030; Distributed Query into pg_merge_job_570015 Executor: Real-Time Task Count: 4 Tasks Shown: All -> Task Node: host=localhost port=57637 dbname=regression -> Aggregate -> Seq Scan on lineitem_290005 lineitem Filter: (l_orderkey > 9030) -> Task Node: host=localhost port=57638 dbname=regression -> Aggregate -> Seq Scan on lineitem_290004 lineitem Filter: (l_orderkey > 9030) -> Task Node: host=localhost port=57637 dbname=regression -> Aggregate -> Seq Scan on lineitem_290007 lineitem Filter: (l_orderkey > 9030) -> Task Node: host=localhost port=57638 dbname=regression -> Aggregate -> Seq Scan on lineitem_290006 lineitem Filter: (l_orderkey > 9030) Master Query -> Aggregate -> Seq Scan on pg_merge_job_570015 SELECT true AS valid FROM explain_xml($$ SELECT avg(l_linenumber) FROM lineitem WHERE l_orderkey > 9030$$); t SELECT true AS valid FROM explain_json($$ SELECT avg(l_linenumber) FROM lineitem WHERE l_orderkey > 9030$$); t -- Test track tracker SET citus.task_executor_type TO 'task-tracker'; SET citus.explain_all_tasks TO off; EXPLAIN (COSTS FALSE) SELECT avg(l_linenumber) FROM lineitem WHERE l_orderkey > 9030; Distributed Query into pg_merge_job_570018 Executor: Task-Tracker Task Count: 4 Tasks Shown: One of 4 -> Task Node: host=localhost port=57637 dbname=regression -> Aggregate -> Seq Scan on lineitem_290005 lineitem Filter: (l_orderkey > 9030) Master Query -> Aggregate -> Seq Scan on pg_merge_job_570018 -- Test re-partition join SET citus.large_table_shard_count TO 1; EXPLAIN (COSTS FALSE) SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey; Distributed Query into pg_merge_job_570021 Executor: Task-Tracker Task Count: 1 Tasks Shown: None, not supported for re-partition queries -> MapMergeJob Map Task Count: 1 Merge Task Count: 1 -> MapMergeJob Map Task Count: 8 Merge Task Count: 1 Master Query -> Aggregate -> Seq Scan on pg_merge_job_570021 EXPLAIN (COSTS FALSE, FORMAT JSON) SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey; [ { "Executor": "Task-Tracker", "Job": { "Task Count": 1, "Tasks Shown": "None, not supported for re-partition queries", "Depended Jobs": [ { "Map Task Count": 1, "Merge Task Count": 1, "Depended Jobs": [ { "Map Task Count": 8, "Merge Task Count": 1 } ] } ] }, "Master Query": [ { "Plan": { "Node Type": "Aggregate", "Strategy": "Plain", "Plans": [ { "Node Type": "Seq Scan", "Parent Relationship": "Outer", "Relation Name": "pg_merge_job_570024", "Alias": "pg_merge_job_570024" } ] } } ] } ] SELECT true AS valid FROM explain_json($$ SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey$$); t EXPLAIN (COSTS FALSE, FORMAT XML) SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey; Task-Tracker 1 None, not supported for re-partition queries 1 1 8 1 Aggregate Plain Seq Scan Outer pg_merge_job_570030 pg_merge_job_570030 SELECT true AS valid FROM explain_xml($$ SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey$$); t EXPLAIN (COSTS FALSE, FORMAT YAML) SELECT count(*) FROM lineitem, orders, customer, supplier WHERE l_orderkey = o_orderkey AND o_custkey = c_custkey AND l_suppkey = s_suppkey; - Executor: "Task-Tracker" Job: Task Count: 1 Tasks Shown: "None, not supported for re-partition queries" Depended Jobs: - Map Task Count: 1 Merge Task Count: 1 Depended Jobs: - Map Task Count: 8 Merge Task Count: 1 Master Query: - Plan: Node Type: "Aggregate" Strategy: "Plain" Plans: - Node Type: "Seq Scan" Parent Relationship: "Outer" Relation Name: "pg_merge_job_570036" Alias: "pg_merge_job_570036"