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

INSERT/SELECT: make SELECT column names unique

pull/3376/head
Hadi Moshayedi 2020-01-16 13:11:31 -08:00
parent 3258d87f3e
commit ce5eea4885
5 changed files with 188 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
src/backend/distributed/executor/insert_select_executor.c
View File

@ -77,6 +77,7 @@ static int PartitionColumnIndex(List *insertTargetList, Var *partitionColumn);
static Expr * CastExpr(Expr *expr, Oid sourceType, Oid targetType, Oid targetCollation, static Expr * CastExpr(Expr *expr, Oid sourceType, Oid targetType, Oid targetCollation,
int targetTypeMod); int targetTypeMod);
static void WrapTaskListForProjection(List *taskList, List *projectedTargetEntries); static void WrapTaskListForProjection(List *taskList, List *projectedTargetEntries);
static void RelableTargetEntryList(List *selectTargetList, List *insertTargetList);
/* /*
@ -155,6 +156,14 @@ CoordinatorInsertSelectExecScanInternal(CustomScanState *node)
selectQuery->targetList, selectQuery->targetList,
targetRelationId); targetRelationId);
/*
* Later we might need to call WrapTaskListForProjection(), which requires
* that select target list has unique names, otherwise the outer query
* cannot select columns unambiguously. So we relabel select columns to
* match target columns.
*/
RelableTargetEntryList(selectQuery->targetList, insertTargetList);
/* /*
* Make a copy of the query, since pg_plan_query may scribble on it and we * Make a copy of the query, since pg_plan_query may scribble on it and we
* want it to be replanned every time if it is stored in a prepared * want it to be replanned every time if it is stored in a prepared
@ -753,6 +762,12 @@ AddInsertSelectCasts(List *insertTargetList, List *selectTargetList,
if (selectEntry->ressortgroupref != 0) if (selectEntry->ressortgroupref != 0)
{ {
/* make sure that the name doesn't match any insert target list entries */
resnameString = makeStringInfo();
appendStringInfo(resnameString, "auto_resjunked_by_citus_%d",
targetEntryIndex);
selectEntry->resname = resnameString->data;
selectEntry->resjunk = true; selectEntry->resjunk = true;
nonProjectedEntries = lappend(nonProjectedEntries, selectEntry); nonProjectedEntries = lappend(nonProjectedEntries, selectEntry);
} }
@ -1077,3 +1092,23 @@ WrapTaskListForProjection(List *taskList, List *projectedTargetEntries)
task->queryString = wrappedQuery->data; task->queryString = wrappedQuery->data;
} }
} }
/*
* RelableTargetEntryList relabels select target list to have matching names with
* insert target list.
*/
static void
RelableTargetEntryList(List *selectTargetList, List *insertTargetList)
{
ListCell *selectTargetCell = NULL;
ListCell *insertTargetCell = NULL;
forboth(selectTargetCell, selectTargetList, insertTargetCell, insertTargetList)
{
TargetEntry *selectTargetEntry = lfirst(selectTargetCell);
TargetEntry *insertTargetEntry = lfirst(insertTargetCell);
selectTargetEntry->resname = insertTargetEntry->resname;
}
}

3
src/test/regress/expected/cte_inline.out
View File

@ -825,9 +825,10 @@ INSERT INTO test_table
FROM FROM
fist_table_cte; fist_table_cte;
DEBUG: distributed INSERT ... SELECT can only select from distributed tables DEBUG: distributed INSERT ... SELECT can only select from distributed tables
DEBUG: Collecting INSERT ... SELECT results on coordinator
DEBUG: CTE fist_table_cte is going to be inlined via distributed planning DEBUG: CTE fist_table_cte is going to be inlined via distributed planning
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'key'
-- the following INSERT..SELECT is even more interesting -- the following INSERT..SELECT is even more interesting
-- the CTE becomes pushdownable -- the CTE becomes pushdownable
INSERT INTO test_table INSERT INTO test_table

110
src/test/regress/expected/insert_select_repartition.out
View File

@ -29,11 +29,11 @@ DETAIL: Subquery contains an operator in the same position as the target table'
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name '?column?' DEBUG: partitioning SELECT query by column index 0 with name 'a'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213585 AS citus_table_alias (a) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_4213583_to_0,repartitioned_results_from_4213584_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213585 AS citus_table_alias (a) SELECT a FROM read_intermediate_results('{repartitioned_results_from_4213583_to_0,repartitioned_results_from_4213584_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213586 AS citus_table_alias (a) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_4213582_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213586 AS citus_table_alias (a) SELECT a FROM read_intermediate_results('{repartitioned_results_from_4213582_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213587 AS citus_table_alias (a) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_4213581_to_2,repartitioned_results_from_4213582_to_2,repartitioned_results_from_4213584_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213587 AS citus_table_alias (a) SELECT a FROM read_intermediate_results('{repartitioned_results_from_4213581_to_2,repartitioned_results_from_4213582_to_2,repartitioned_results_from_4213584_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213588 AS citus_table_alias (a) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_4213581_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213588 AS citus_table_alias (a) SELECT a FROM read_intermediate_results('{repartitioned_results_from_4213581_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer)
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table WHERE a=-1 OR a=-3 OR a=-7 ORDER BY a; SELECT * FROM target_table WHERE a=-1 OR a=-3 OR a=-7 ORDER BY a;
a a
@ -83,9 +83,9 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 2 with name 'mapped_key' DEBUG: partitioning SELECT query by column index 2 with name 'key'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, value, key) SELECT f1, value, mapped_key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, mapped_key insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, value, key) SELECT f1, value, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, key insert_select_repartition.composite_key_type)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, value, key) SELECT f1, value, mapped_key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, mapped_key insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, value, key) SELECT f1, value, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, key insert_select_repartition.composite_key_type)
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY key; SELECT * FROM target_table ORDER BY key;
f1 | value | key f1 | value | key
@ -113,9 +113,9 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 2 with name 'mapped_key' DEBUG: partitioning SELECT query by column index 2 with name 'key'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, value, key) SELECT f1, value, mapped_key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, mapped_key insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, value, key) SELECT f1, value, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, key insert_select_repartition.composite_key_type)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, value, key) SELECT f1, value, mapped_key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, mapped_key insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, value, key) SELECT f1, value, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, value integer, key insert_select_repartition.composite_key_type)
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY key; SELECT * FROM target_table ORDER BY key;
f1 | value | key f1 | value | key
@ -137,9 +137,9 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 1 with name 'key_renamed' DEBUG: partitioning SELECT query by column index 1 with name 'key'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, key) SELECT f1, key_renamed FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key_renamed insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, key) SELECT f1, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0,repartitioned_results_from_4213590_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key insert_select_repartition.composite_key_type)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, key) SELECT f1, key_renamed FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key_renamed insert_select_repartition.composite_key_type) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, key) SELECT f1, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key insert_select_repartition.composite_key_type)
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY key; SELECT * FROM target_table ORDER BY key;
f1 | value | key f1 | value | key
@ -163,9 +163,9 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 1 with name 'key_renamed' DEBUG: partitioning SELECT query by column index 1 with name 'key'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, key) SELECT f1, key_renamed FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key_renamed insert_select_repartition.composite_key_type) ON CONFLICT(key) DO UPDATE SET f1 = 1 DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213591 AS citus_table_alias (f1, key) SELECT f1, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_0}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key insert_select_repartition.composite_key_type) ON CONFLICT(key) DO UPDATE SET f1 = 1
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, key) SELECT f1, key_renamed FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key_renamed insert_select_repartition.composite_key_type) ON CONFLICT(key) DO UPDATE SET f1 = 1 DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213592 AS citus_table_alias (f1, key) SELECT f1, key FROM read_intermediate_results('{repartitioned_results_from_4213589_to_1,repartitioned_results_from_4213590_to_1}'::text[], 'text'::citus_copy_format) intermediate_result(f1 integer, key insert_select_repartition.composite_key_type) ON CONFLICT(key) DO UPDATE SET f1 = 1
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY key; SELECT * FROM target_table ORDER BY key;
f1 | value | key f1 | value | key
@ -213,9 +213,9 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The data type of the target table's partition column should exactly match the data type of the corresponding simple column reference in the subquery. DETAIL: The data type of the target table's partition column should exactly match the data type of the corresponding simple column reference in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'col_1'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213593 AS citus_table_alias (col_1, col_2) SELECT auto_coerced_by_citus_0, auto_coerced_by_citus_1 FROM read_intermediate_results('{repartitioned_results_from_4213597_to_0,repartitioned_results_from_4213600_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer, auto_coerced_by_citus_1 integer) ON CONFLICT(col_1) DO UPDATE SET col_2 = excluded.col_2 DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213593 AS citus_table_alias (col_1, col_2) SELECT col_1, col_2 FROM read_intermediate_results('{repartitioned_results_from_4213597_to_0,repartitioned_results_from_4213600_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(col_1 integer, col_2 integer) ON CONFLICT(col_1) DO UPDATE SET col_2 = excluded.col_2
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213594 AS citus_table_alias (col_1, col_2) SELECT auto_coerced_by_citus_0, auto_coerced_by_citus_1 FROM read_intermediate_results('{repartitioned_results_from_4213599_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer, auto_coerced_by_citus_1 integer) ON CONFLICT(col_1) DO UPDATE SET col_2 = excluded.col_2 DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213594 AS citus_table_alias (col_1, col_2) SELECT col_1, col_2 FROM read_intermediate_results('{repartitioned_results_from_4213599_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(col_1 integer, col_2 integer) ON CONFLICT(col_1) DO UPDATE SET col_2 = excluded.col_2
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY 1; SELECT * FROM target_table ORDER BY 1;
col_1 | col_2 col_1 | col_2
@ -506,7 +506,7 @@ INSERT INTO target_table
SELECT mapped_key, c FROM t NATURAL JOIN source_table; SELECT mapped_key, c FROM t NATURAL JOIN source_table;
DEBUG: volatile functions are not allowed in distributed INSERT ... SELECT queries DEBUG: volatile functions are not allowed in distributed INSERT ... SELECT queries
DEBUG: generating subplan XXX_1 for CTE t: SELECT mapped_key, a, c FROM insert_select_repartition.source_table WHERE ((a)::double precision OPERATOR(pg_catalog.>) floor(random())) DEBUG: generating subplan XXX_1 for CTE t: SELECT mapped_key, a, c FROM insert_select_repartition.source_table WHERE ((a)::double precision OPERATOR(pg_catalog.>) floor(random()))
DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT mapped_key, (c)::integer[] AS auto_coerced_by_citus_1 FROM (SELECT t.mapped_key, t.c FROM ((SELECT intermediate_result.mapped_key, intermediate_result.a, intermediate_result.c FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(mapped_key integer, a integer, c double precision[])) t JOIN insert_select_repartition.source_table USING (mapped_key, a, c))) citus_insert_select_subquery DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT mapped_key AS a, (c)::integer[] AS b FROM (SELECT t.mapped_key, t.c FROM ((SELECT intermediate_result.mapped_key, intermediate_result.a, intermediate_result.c FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(mapped_key integer, a integer, c double precision[])) t JOIN insert_select_repartition.source_table USING (mapped_key, a, c))) citus_insert_select_subquery
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY a; SELECT * FROM target_table ORDER BY a;
@ -573,6 +573,32 @@ EXPLAIN INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
-> Seq Scan on source_table_4213606 source_table (cost=0.00..32.60 rows=2260 width=8) -> Seq Scan on source_table_4213606 source_table (cost=0.00..32.60 rows=2260 width=8)
(10 rows) (10 rows)
--
-- Duplicate names in target list
--
TRUNCATE target_table;
SET client_min_messages TO DEBUG2;
INSERT INTO target_table
SELECT max(b), max(b) FROM source_table GROUP BY a;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'a'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213610 AS citus_table_alias (a, b) SELECT a, b FROM read_intermediate_results('{repartitioned_results_from_4213606_to_0,repartitioned_results_from_4213607_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213611 AS citus_table_alias (a, b) SELECT a, b FROM read_intermediate_results('{repartitioned_results_from_4213607_to_1,repartitioned_results_from_4213609_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213612 AS citus_table_alias (a, b) SELECT a, b FROM read_intermediate_results('{repartitioned_results_from_4213606_to_2,repartitioned_results_from_4213607_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer)
RESET client_min_messages;
SELECT * FROM target_table ORDER BY a;
a | b
---------------------------------------------------------------------
9 | 9
49 | 49
121 | 121
225 | 225
361 | 361
400 | 400
(6 rows)
-- --
-- Prepared INSERT/SELECT -- Prepared INSERT/SELECT
-- --
@ -726,8 +752,8 @@ INSERT INTO target_table SELECT 1.12, b::bigint FROM source_table WHERE b IS NOT
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'a'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213617 AS citus_table_alias (a, b) SELECT auto_coerced_by_citus_0, auto_coerced_by_citus_1 FROM read_intermediate_results('{repartitioned_results_from_4213613_to_0,repartitioned_results_from_4213614_to_0,repartitioned_results_from_4213615_to_0,repartitioned_results_from_4213616_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer, auto_coerced_by_citus_1 integer) DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213617 AS citus_table_alias (a, b) SELECT a, b FROM read_intermediate_results('{repartitioned_results_from_4213613_to_0,repartitioned_results_from_4213614_to_0,repartitioned_results_from_4213615_to_0,repartitioned_results_from_4213616_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer)
RESET client_min_messages; RESET client_min_messages;
SELECT * FROM target_table ORDER BY a, b; SELECT * FROM target_table ORDER BY a, b;
a | b a | b
@ -811,6 +837,44 @@ SELECT * FROM run_command_on_placements('target_table', 'select count(*) from %s
localhost | 57638 | 4213628 | t | 2 localhost | 57638 | 4213628 | t | 2
(8 rows) (8 rows)
DROP TABLE source_table, target_table;
--
-- Select column names should be unique
--
SET citus.shard_replication_factor TO 1;
SET citus.shard_count TO 4;
CREATE TABLE source_table(a int, b int);
SELECT create_distributed_table('source_table', 'a');
create_distributed_table
---------------------------------------------------------------------
(1 row)
SET citus.shard_count TO 3;
CREATE TABLE target_table(a int, b int, c int, d int, e int, f int);
SELECT create_distributed_table('target_table', 'a');
create_distributed_table
---------------------------------------------------------------------
(1 row)
INSERT INTO source_table SELECT i, i * i FROM generate_series(1, 10) i;
SET client_min_messages TO DEBUG2;
INSERT INTO target_table SELECT a AS aa, b AS aa, 1 AS aa, 2 AS aa FROM source_table;
DEBUG: INSERT target table and the source relation of the SELECT partition column value must be colocated in distributed INSERT ... SELECT
DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'a'
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213633 AS citus_table_alias (a, b, c, d) SELECT a, b, c, d FROM read_intermediate_results('{repartitioned_results_from_4213629_to_0,repartitioned_results_from_4213630_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer, c integer, d integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213634 AS citus_table_alias (a, b, c, d) SELECT a, b, c, d FROM read_intermediate_results('{repartitioned_results_from_4213630_to_1,repartitioned_results_from_4213631_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer, c integer, d integer)
DEBUG: distributed statement: INSERT INTO insert_select_repartition.target_table_4213635 AS citus_table_alias (a, b, c, d) SELECT a, b, c, d FROM read_intermediate_results('{repartitioned_results_from_4213632_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(a integer, b integer, c integer, d integer)
RESET client_min_messages;
SELECT count(*) FROM target_table;
count
---------------------------------------------------------------------
10
(1 row)
DROP TABLE source_table, target_table; DROP TABLE source_table, target_table;
SET client_min_messages TO WARNING; SET client_min_messages TO WARNING;
DROP SCHEMA insert_select_repartition CASCADE; DROP SCHEMA insert_select_repartition CASCADE;

56
src/test/regress/expected/multi_insert_select.out
View File

@ -696,7 +696,7 @@ DEBUG: distributed INSERT ... SELECT can only select from distributed tables
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: generating subplan XXX_1 for CTE fist_table_agg: SELECT (max(value_1) OPERATOR(pg_catalog.+) 1) AS v1_agg, user_id FROM public.raw_events_first GROUP BY user_id DEBUG: generating subplan XXX_1 for CTE fist_table_agg: SELECT (max(value_1) OPERATOR(pg_catalog.+) 1) AS v1_agg, user_id FROM public.raw_events_first GROUP BY user_id
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT user_id, v1_agg FROM (SELECT fist_table_agg.v1_agg, fist_table_agg.user_id FROM (SELECT intermediate_result.v1_agg, intermediate_result.user_id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v1_agg integer, user_id integer)) fist_table_agg) citus_insert_select_subquery DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT user_id, v1_agg AS value_1_agg FROM (SELECT fist_table_agg.v1_agg, fist_table_agg.user_id FROM (SELECT intermediate_result.v1_agg, intermediate_result.user_id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v1_agg integer, user_id integer)) fist_table_agg) citus_insert_select_subquery
DEBUG: Creating router plan DEBUG: Creating router plan
DEBUG: Plan is router executable DEBUG: Plan is router executable
DEBUG: Collecting INSERT ... SELECT results on coordinator DEBUG: Collecting INSERT ... SELECT results on coordinator
@ -1083,7 +1083,7 @@ DEBUG: join prunable for intervals [1073741824,2147483647] and [-2147483648,-10
DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1] DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1]
DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823] DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823]
DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_3 DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_3
DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT int4(id) AS auto_coerced_by_citus_0, int4(v1) AS auto_coerced_by_citus_1, int8(v4) AS auto_coerced_by_citus_2 FROM (SELECT intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 bigint, id double precision)) foo DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT int4(id) AS user_id, int4(v1) AS value_1_agg, int8(v4) AS value_4_agg FROM (SELECT intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 bigint, id double precision)) foo
DEBUG: Creating router plan DEBUG: Creating router plan
DEBUG: Plan is router executable DEBUG: Plan is router executable
DEBUG: Collecting INSERT ... SELECT results on coordinator DEBUG: Collecting INSERT ... SELECT results on coordinator
@ -1116,7 +1116,7 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'value_1' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
ERROR: the partition column value cannot be NULL ERROR: the partition column value cannot be NULL
CONTEXT: while executing command on localhost:xxxxx CONTEXT: while executing command on localhost:xxxxx
INSERT INTO raw_events_second INSERT INTO raw_events_second
@ -1128,11 +1128,11 @@ DETAIL: Subquery contains an operator in the same position as the target table'
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name '?column?' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300004 AS citus_table_alias (user_id) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_13300000_to_0,repartitioned_results_from_13300001_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300004 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300000_to_0,repartitioned_results_from_13300001_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_13300000_to_1,repartitioned_results_from_13300001_to_1,repartitioned_results_from_13300003_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300000_to_1,repartitioned_results_from_13300001_to_1,repartitioned_results_from_13300003_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300006 AS citus_table_alias (user_id) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_13300001_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300006 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300001_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300007 AS citus_table_alias (user_id) SELECT "?column?" FROM read_intermediate_results('{repartitioned_results_from_13300000_to_3,repartitioned_results_from_13300002_to_3,repartitioned_results_from_13300003_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result("?column?" integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300007 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300000_to_3,repartitioned_results_from_13300002_to_3,repartitioned_results_from_13300003_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
INSERT INTO raw_events_second INSERT INTO raw_events_second
(user_id) (user_id)
SELECT user_id :: bigint SELECT user_id :: bigint
@ -1142,11 +1142,11 @@ DETAIL: Subquery contains an explicit cast in the same position as the target t
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300004 AS citus_table_alias (user_id) SELECT auto_coerced_by_citus_0 FROM read_intermediate_results('{repartitioned_results_from_13300000_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300004 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300000_to_0}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id) SELECT auto_coerced_by_citus_0 FROM read_intermediate_results('{repartitioned_results_from_13300001_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300005 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300001_to_1}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300006 AS citus_table_alias (user_id) SELECT auto_coerced_by_citus_0 FROM read_intermediate_results('{repartitioned_results_from_13300002_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300006 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300002_to_2}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300007 AS citus_table_alias (user_id) SELECT auto_coerced_by_citus_0 FROM read_intermediate_results('{repartitioned_results_from_13300003_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result(auto_coerced_by_citus_0 integer) DEBUG: distributed statement: INSERT INTO public.raw_events_second_13300007 AS citus_table_alias (user_id) SELECT user_id FROM read_intermediate_results('{repartitioned_results_from_13300003_to_3}'::text[], 'binary'::citus_copy_format) intermediate_result(user_id integer)
INSERT INTO agg_events INSERT INTO agg_events
(value_3_agg, (value_3_agg,
value_4_agg, value_4_agg,
@ -1165,7 +1165,7 @@ DETAIL: Subquery contains an aggregation in the same position as the target tab
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
ERROR: the partition column value cannot be NULL ERROR: the partition column value cannot be NULL
CONTEXT: while executing command on localhost:xxxxx CONTEXT: while executing command on localhost:xxxxx
INSERT INTO agg_events INSERT INTO agg_events
@ -1186,7 +1186,7 @@ DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition
DETAIL: The target table's partition column should correspond to a partition column in the subquery. DETAIL: The target table's partition column should correspond to a partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'value_2' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
ERROR: the partition column value cannot be NULL ERROR: the partition column value cannot be NULL
CONTEXT: while executing command on localhost:xxxxx CONTEXT: while executing command on localhost:xxxxx
-- tables should be co-located -- tables should be co-located
@ -1240,10 +1240,10 @@ DEBUG: join prunable for intervals [1073741824,2147483647] and [-2147483648,-10
DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1] DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1]
DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823] DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823]
DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_1 HAVING (sum(raw_events_second.value_4) OPERATOR(pg_catalog.>) (10)::numeric) DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_1 HAVING (sum(raw_events_second.value_4) OPERATOR(pg_catalog.>) (10)::numeric)
DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT int4(f2.id) AS auto_coerced_by_citus_0 FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first, public.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 intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 integer, id bigint)) foo2) f2 ON ((f.id OPERATOR(pg_catalog.=) f2.id))) DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT int4(f2.id) AS user_id FROM ((SELECT foo.id FROM (SELECT reference_table.user_id AS id FROM public.raw_events_first, public.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 intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 integer, id bigint)) foo2) f2 ON ((f.id OPERATOR(pg_catalog.=) f2.id)))
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
-- the second part of the query is not routable since -- the second part of the query is not routable since
-- GROUP BY not on the partition column (i.e., value_1) and thus join -- GROUP BY not on the partition column (i.e., value_1) and thus join
-- on f.id = f2.id is not on the partition key (instead on the sum of partition key) -- on f.id = f2.id is not on the partition key (instead on the sum of partition key)
@ -1286,10 +1286,10 @@ DEBUG: join prunable for intervals [1073741824,2147483647] and [-2147483648,-10
DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1] DEBUG: join prunable for intervals [1073741824,2147483647] and [-1073741824,-1]
DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823] DEBUG: join prunable for intervals [1073741824,2147483647] and [0,1073741823]
DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_1 HAVING (sum(raw_events_second.value_4) OPERATOR(pg_catalog.>) (10)::numeric) DEBUG: generating subplan XXX_1 for subquery 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 public.raw_events_first, public.raw_events_second WHERE (raw_events_first.user_id OPERATOR(pg_catalog.=) raw_events_second.user_id) GROUP BY raw_events_second.value_1 HAVING (sum(raw_events_second.value_4) OPERATOR(pg_catalog.>) (10)::numeric)
DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT f.id FROM ((SELECT foo.id FROM (SELECT raw_events_first.user_id AS id FROM public.raw_events_first, public.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 intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 integer, id bigint)) foo2) f2 ON ((f.id OPERATOR(pg_catalog.=) f2.id))) DEBUG: Plan XXX query after replacing subqueries and CTEs: SELECT f.id AS user_id FROM ((SELECT foo.id FROM (SELECT raw_events_first.user_id AS id FROM public.raw_events_first, public.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 intermediate_result.v4, intermediate_result.v1, intermediate_result.id FROM read_intermediate_result('XXX_1'::text, 'binary'::citus_copy_format) intermediate_result(v4 numeric, v1 integer, id bigint)) foo2) f2 ON ((f.id OPERATOR(pg_catalog.=) f2.id)))
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'id' DEBUG: partitioning SELECT query by column index 0 with name 'user_id'
-- cannot pushdown the query since the JOIN is not equi JOIN -- cannot pushdown the query since the JOIN is not equi JOIN
INSERT INTO agg_events INSERT INTO agg_events
(user_id, value_4_agg) (user_id, value_4_agg)
@ -2114,42 +2114,42 @@ DETAIL: Subquery contains a case expression in the same position as the target
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'case' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT COALESCE(part_col, 'onder') FROM text_table; INSERT INTO text_table (part_col) SELECT COALESCE(part_col, 'onder') FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains a coalesce expression in the same position as the target table's partition column. DETAIL: Subquery contains a coalesce expression in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'coalesce' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT GREATEST(part_col, 'jason') FROM text_table; INSERT INTO text_table (part_col) SELECT GREATEST(part_col, 'jason') FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains a min/max expression in the same position as the target table's partition column. DETAIL: Subquery contains a min/max expression in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'greatest' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT LEAST(part_col, 'andres') FROM text_table; INSERT INTO text_table (part_col) SELECT LEAST(part_col, 'andres') FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains a min/max expression in the same position as the target table's partition column. DETAIL: Subquery contains a min/max expression in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'least' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT NULLIF(part_col, 'metin') FROM text_table; INSERT INTO text_table (part_col) SELECT NULLIF(part_col, 'metin') FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains an expression that is not a simple column reference in the same position as the target table's partition column. DETAIL: Subquery contains an expression that is not a simple column reference in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'nullif' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT part_col isnull FROM text_table; INSERT INTO text_table (part_col) SELECT part_col isnull FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains an expression that is not a simple column reference in the same position as the target table's partition column. DETAIL: Subquery contains an expression that is not a simple column reference in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT part_col::text from char_table; INSERT INTO text_table (part_col) SELECT part_col::text from char_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains an explicit coercion in the same position as the target table's partition column. DETAIL: Subquery contains an explicit coercion in the same position as the target table's partition column.
@ -2163,20 +2163,20 @@ DETAIL: Subquery contains an expression that is not a simple column reference i
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT val FROM text_table; INSERT INTO text_table (part_col) SELECT val FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: The data type of the target table's partition column should exactly match the data type of the corresponding simple column reference in the subquery. DETAIL: The data type of the target table's partition column should exactly match the data type of the corresponding simple column reference in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'auto_coerced_by_citus_0' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
INSERT INTO text_table (part_col) SELECT val::text FROM text_table; INSERT INTO text_table (part_col) SELECT val::text FROM text_table;
DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match DEBUG: cannot perform distributed INSERT INTO ... SELECT because the partition columns in the source table and subquery do not match
DETAIL: Subquery contains an explicit coercion in the same position as the target table's partition column. DETAIL: Subquery contains an explicit coercion in the same position as the target table's partition column.
HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery. HINT: Ensure the target table's partition column has a corresponding simple column reference to a distributed table's partition column in the subquery.
DEBUG: Router planner cannot handle multi-shard select queries DEBUG: Router planner cannot handle multi-shard select queries
DEBUG: performing repartitioned INSERT ... SELECT DEBUG: performing repartitioned INSERT ... SELECT
DEBUG: partitioning SELECT query by column index 0 with name 'val' DEBUG: partitioning SELECT query by column index 0 with name 'part_col'
RESET client_min_messages; RESET client_min_messages;
insert into table_with_starts_with_defaults (b,c) select b,c FROM table_with_starts_with_defaults; 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

36
src/test/regress/sql/insert_select_repartition.sql
View File

@ -271,6 +271,19 @@ SELECT * FROM target_table ORDER BY a;
EXPLAIN INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a; EXPLAIN INSERT INTO target_table SELECT a, max(b) FROM source_table GROUP BY a;
--
-- Duplicate names in target list
--
TRUNCATE target_table;
SET client_min_messages TO DEBUG2;
INSERT INTO target_table
SELECT max(b), max(b) FROM source_table GROUP BY a;
RESET client_min_messages;
SELECT * FROM target_table ORDER BY a;
-- --
-- Prepared INSERT/SELECT -- Prepared INSERT/SELECT
-- --
@ -384,5 +397,28 @@ SELECT * FROM run_command_on_placements('target_table', 'select count(*) from %s
DROP TABLE source_table, target_table; DROP TABLE source_table, target_table;
--
-- Select column names should be unique
--
SET citus.shard_replication_factor TO 1;
SET citus.shard_count TO 4;
CREATE TABLE source_table(a int, b int);
SELECT create_distributed_table('source_table', 'a');
SET citus.shard_count TO 3;
CREATE TABLE target_table(a int, b int, c int, d int, e int, f int);
SELECT create_distributed_table('target_table', 'a');
INSERT INTO source_table SELECT i, i * i FROM generate_series(1, 10) i;
SET client_min_messages TO DEBUG2;
INSERT INTO target_table SELECT a AS aa, b AS aa, 1 AS aa, 2 AS aa FROM source_table;
RESET client_min_messages;
SELECT count(*) FROM target_table;
DROP TABLE source_table, target_table;
SET client_min_messages TO WARNING; SET client_min_messages TO WARNING;
DROP SCHEMA insert_select_repartition CASCADE; DROP SCHEMA insert_select_repartition CASCADE;