external
/
citus
mirror of https://github.com/citusdata/citus.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Increate subquery test coverage

pull/896/head
Onder Kalaci 2016-10-20 13:13:48 +03:00
parent 3a7e10a9b1
commit 3ca181fea8
2 changed files with 574 additions and 171 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

438
src/test/regress/expected/multi_insert_select.out
View File

@ -539,17 +539,181 @@ DEBUG: Plan is router executable
ERROR: duplicate key value violates unique constraint "agg_events_user_id_value_1_agg_key_13300008"
DETAIL: Key (user_id, value_1_agg)=(1, 10) already exists.
CONTEXT: while executing command on localhost:57638
-- join between subqueries
INSERT INTO agg_events
(user_id)
SELECT f2.id FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300008 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_13300000 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300000 raw_events_first, public.raw_events_second_13300004 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300009 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_13300001 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300001 raw_events_first, public.raw_events_second_13300005 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300010 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_13300002 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300002 raw_events_first, public.raw_events_second_13300006 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
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 = 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 = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))
DEBUG: ProcessQuery
DEBUG: Plan is router executable
DEBUG: CommitTransactionCommand
DEBUG: CommitTransaction
DEBUG: name: unnamed; blockState: STARTED; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
DEBUG: sent COMMIT over connection 13300008
DEBUG: sent COMMIT over connection 13300008
DEBUG: sent COMMIT over connection 13300011
DEBUG: sent COMMIT over connection 13300011
DEBUG: sent COMMIT over connection 13300009
DEBUG: sent COMMIT over connection 13300009
DEBUG: sent COMMIT over connection 13300010
DEBUG: sent COMMIT over connection 13300010
-- add one more level subqueris on top of subquery JOINs
INSERT INTO agg_events
(user_id, value_4_agg)
SELECT
outer_most.id, max(outer_most.value)
FROM
(
SELECT f2.id as id, f2.v4 as value FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id)) as outer_most
GROUP BY
outer_most.id;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300008 AS citus_table_alias (user_id, value_4_agg) SELECT id, max(value) AS max FROM (SELECT f2.id, f2.v4 AS value FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first_13300000 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300000 raw_events_first, public.raw_events_second_13300004 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))) outer_most GROUP BY id
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300009 AS citus_table_alias (user_id, value_4_agg) SELECT id, max(value) AS max FROM (SELECT f2.id, f2.v4 AS value FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first_13300001 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300001 raw_events_first, public.raw_events_second_13300005 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))) outer_most GROUP BY id
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300007
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300003
DEBUG: distributed statement: INSERT INTO public.agg_events_13300010 AS citus_table_alias (user_id, value_4_agg) SELECT id, max(value) AS max FROM (SELECT f2.id, f2.v4 AS value FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first_13300002 raw_events_first, public.reference_table_13300012 reference_table WHERE (raw_events_first.user_id = 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_13300002 raw_events_first, public.raw_events_second_13300006 raw_events_second WHERE (raw_events_first.user_id = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))) outer_most GROUP BY id
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300004
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300000
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: distributed statement: INSERT INTO public.agg_events_13300011 AS citus_table_alias (user_id, value_4_agg) SELECT id, max(value) AS max FROM (SELECT f2.id, f2.v4 AS value 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 = 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 = raw_events_second.user_id) GROUP BY raw_events_second.user_id HAVING (sum(raw_events_second.value_4) > (10)::numeric)) foo2) f2 ON ((f.id = f2.id)))) outer_most GROUP BY id
DEBUG: ProcessQuery
DEBUG: Plan is router executable
DEBUG: CommitTransactionCommand
DEBUG: CommitTransaction
DEBUG: name: unnamed; blockState: STARTED; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
DEBUG: sent COMMIT over connection 13300008
DEBUG: sent COMMIT over connection 13300008
DEBUG: sent COMMIT over connection 13300011
DEBUG: sent COMMIT over connection 13300011
DEBUG: sent COMMIT over connection 13300009
DEBUG: sent COMMIT over connection 13300009
DEBUG: sent COMMIT over connection 13300010
DEBUG: sent COMMIT over connection 13300010
-- some UPSERTS
INSERT INTO agg_events AS ae
(
user_id,
value_1_agg,
agg_time
(
user_id,
value_1_agg,
agg_time
)
SELECT user_id,
value_1,
time
FROM raw_events_first
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
@ -822,124 +986,224 @@ DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid:
ERROR: cannot perform distributed planning for the given modification
DETAIL: Set operations are not allowed in INSERT ... SELECT queries
-- unsupported JOIN
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;
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;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: cannot perform distributed planning for the given modification
DETAIL: Select query cannot be pushed down to the worker.
-- INSERT partition column does not match with SELECT partition column
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.value_3 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.value_3) AS foo;
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.value_3 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.value_3) AS foo;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
-- error cases
-- no part column at all
INSERT INTO raw_events_second
(value_1)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(value_1)
SELECT value_1
FROM raw_events_first;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO raw_events_second
(value_1)
SELECT user_id
FROM raw_events_first;
INSERT INTO raw_events_second
(value_1)
SELECT user_id
FROM raw_events_first;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO raw_events_second
(user_id)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT value_1
FROM raw_events_first;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO raw_events_second
(user_id)
SELECT user_id * 2
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id * 2
FROM raw_events_first;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO raw_events_second
(user_id)
SELECT user_id :: bigint
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id :: bigint
FROM raw_events_first;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
Avg(value_2)
FROM raw_events_first
GROUP BY user_id;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
Avg(value_2)
FROM raw_events_first
GROUP BY user_id;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
value_2
FROM raw_events_first
GROUP BY user_id,
value_2;
value_2
FROM raw_events_first
GROUP BY user_id,
value_2;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
-- tables should be co-located
INSERT INTO agg_events (user_id) SELECT user_id FROM reference_table;
INSERT INTO agg_events (user_id)
SELECT
user_id
FROM
reference_table;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: INSERT target table and the source relation of the SELECT partition column value must be colocated
-- unsupported joins between subqueries
-- we do not return bare partition column on the inner query
INSERT INTO agg_events
(user_id)
SELECT f2.id FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(SELECT v4,
v1,
id
FROM (SELECT SUM(raw_events_second.value_4) AS v4,
raw_events_second.value_1 AS v1,
SUM(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.value_1
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: SELECT query should return bare partition column on the same ordinal position with INSERT query's partition column
-- the second part of the query is not routable since
-- no GROUP BY on the partition column
INSERT INTO agg_events
(user_id)
SELECT f.id FROM
(SELECT
id
FROM (SELECT raw_events_first.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(SELECT v4,
v1,
id
FROM (SELECT SUM(raw_events_second.value_4) AS v4,
raw_events_second.value_1 AS v1,
SUM(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.value_1
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
ERROR: cannot perform distributed planning for the given modification
DETAIL: Select query cannot be pushed down to the worker.
-- cannot pushdown the query since the JOIN is not equi JOIN
INSERT INTO agg_events
(user_id, value_4_agg)
SELECT
outer_most.id, max(outer_most.value)
FROM
(
SELECT f2.id as id, f2.v4 as value FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id != f2.id)) as outer_most
GROUP BY outer_most.id;
DEBUG: StartTransactionCommand
DEBUG: StartTransaction
DEBUG: name: unnamed; blockState: DEFAULT; state: INPROGR, xid/subid/cid: 0/1/0, nestlvl: 1, children:
DEBUG: predicate pruning for shardId 13300001
DEBUG: predicate pruning for shardId 13300002
DEBUG: predicate pruning for shardId 13300003
DEBUG: predicate pruning for shardId 13300005
DEBUG: predicate pruning for shardId 13300006
DEBUG: predicate pruning for shardId 13300007
ERROR: cannot perform distributed planning for the given modification
DETAIL: Select query cannot be pushed down to the worker.

307
src/test/regress/sql/multi_insert_select.sql
View File

@ -180,18 +180,72 @@ FROM (SELECT SUM(raw_events_second.value_4) AS v4,
WHERE raw_events_first.user_id = raw_events_second.user_id
GROUP BY raw_events_second.user_id) AS foo;
-- join between subqueries
INSERT INTO agg_events
(user_id)
SELECT f2.id FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
-- add one more level subqueris on top of subquery JOINs
INSERT INTO agg_events
(user_id, value_4_agg)
SELECT
outer_most.id, max(outer_most.value)
FROM
(
SELECT f2.id as id, f2.v4 as value FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id)) as outer_most
GROUP BY
outer_most.id;
-- some UPSERTS
INSERT INTO agg_events AS ae
(
user_id,
value_1_agg,
agg_time
(
user_id,
value_1_agg,
agg_time
)
SELECT user_id,
value_1,
time
FROM raw_events_first
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
@ -287,93 +341,178 @@ FROM
(SELECT user_id FROM raw_events_second where user_id = 17)) as foo;
-- unsupported JOIN
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;
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;
-- INSERT partition column does not match with SELECT partition column
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.value_3 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.value_3) AS foo;
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.value_3 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.value_3) AS foo;
-- error cases
-- no part column at all
INSERT INTO raw_events_second
(value_1)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(value_1)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(value_1)
SELECT user_id
FROM raw_events_first;
INSERT INTO raw_events_second
(value_1)
SELECT user_id
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT value_1
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id * 2
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id * 2
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id :: bigint
FROM raw_events_first;
INSERT INTO raw_events_second
(user_id)
SELECT user_id :: bigint
FROM raw_events_first;
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
Avg(value_2)
FROM raw_events_first
GROUP BY user_id;
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
Avg(value_2)
FROM raw_events_first
GROUP BY user_id;
INSERT INTO agg_events
(value_3_agg,
value_4_agg,
value_1_agg,
value_2_agg,
user_id)
SELECT SUM(value_3),
Count(value_4),
user_id,
SUM(value_1),
value_2
FROM raw_events_first
GROUP BY user_id,
value_2;
value_2
FROM raw_events_first
GROUP BY user_id,
value_2;
-- tables should be co-located
INSERT INTO agg_events (user_id) SELECT user_id FROM reference_table;
INSERT INTO agg_events (user_id)
SELECT
user_id
FROM
reference_table;
-- unsupported joins between subqueries
-- we do not return bare partition column on the inner query
INSERT INTO agg_events
(user_id)
SELECT f2.id FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(SELECT v4,
v1,
id
FROM (SELECT SUM(raw_events_second.value_4) AS v4,
raw_events_second.value_1 AS v1,
SUM(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.value_1
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
-- the second part of the query is not routable since
-- no GROUP BY on the partition column
INSERT INTO agg_events
(user_id)
SELECT f.id FROM
(SELECT
id
FROM (SELECT raw_events_first.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(SELECT v4,
v1,
id
FROM (SELECT SUM(raw_events_second.value_4) AS v4,
raw_events_second.value_1 AS v1,
SUM(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.value_1
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id = f2.id);
-- cannot pushdown the query since the JOIN is not equi JOIN
INSERT INTO agg_events
(user_id, value_4_agg)
SELECT
outer_most.id, max(outer_most.value)
FROM
(
SELECT f2.id as id, f2.v4 as value FROM
(SELECT
id
FROM (SELECT reference_table.user_id AS id
FROM raw_events_first,
reference_table
WHERE raw_events_first.user_id = reference_table.user_id ) AS foo) as f
INNER JOIN
(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
HAVING SUM(raw_events_second.value_4) > 10) AS foo2 ) as f2
ON (f.id != f2.id)) as outer_most
GROUP BY outer_most.id;