Skip to content

Fix ORCA direct dispatch for multi-value quals #1852

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
zhangwenchao-123 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Fix ORCA direct dispatch for multi-value quals #1852

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 13 additions & 13 deletions contrib/pax_storage/src/test/regress/expected/bfv_dd_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,7 +88,7 @@ INFO: (slice 1) Dispatch command to SINGLE content

-- single column distr key
select * from dd_singlecol_1 where a in (10,11,12);
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
a | b
----+----
10 | 10
Expand All @@ -97,7 +97,7 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
(3 rows)

select * from dd_singlecol_1 where a=10 or a=11 or a=12;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
a | b
----+----
10 | 10
Expand All @@ -106,7 +106,7 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
(3 rows)

select * from dd_singlecol_1 where a is null or a=1;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b
---+---
1 | 1
Expand All @@ -115,7 +115,7 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2

-- projections and disjunction
select b from dd_singlecol_1 where a=1 or a=2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
b
---
1
Expand Down Expand Up @@ -175,7 +175,7 @@ INFO: (slice 0) Dispatch command to SINGLE content
INFO: Distributed transaction command 'Distributed Commit (one-phase)' to SINGLE content
-- disjunction with partitioned tables
select * from dd_part_singlecol where a in (10,11,12);
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
a | b | c
----+----+----
10 | 20 | 30
Expand All @@ -184,7 +184,7 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
(3 rows)

select * from dd_part_singlecol where a=10 or a=11 or a=12;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
a | b | c
----+----+----
10 | 20 | 30
Expand All @@ -193,7 +193,7 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
(3 rows)

select * from dd_part_singlecol where a is null or a=1;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b | c
---+---+---
| |
Expand Down Expand Up @@ -321,21 +321,21 @@ analyze dd_singlecol_idx2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
-- disjunction with index scans
select * from dd_singlecol_idx where (a=1 or a=2) and b<2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b | c
---+---+---
1 | 1 | 1
(1 row)

select 'one' from dd_singlecol_idx where (a=1 or a=2) and b=1;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
?column?
----------
one
(1 row)

select a, count(*) from dd_singlecol_idx where (a=1 or a=2) and b=1 group by a;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | count
---+-------
1 | 1
Expand Down Expand Up @@ -369,21 +369,21 @@ analyze dd_singlecol_bitmap_idx;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
-- disjunction with bitmap index scans
select * from dd_singlecol_bitmap_idx where (a=1 or a=2) and b<2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b | c
---+---+---
1 | 1 | 1
(1 row)

select * from dd_singlecol_bitmap_idx where (a=1 or a=2) and b=2 and c=2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b | c
---+---+---
2 | 2 | 2
(1 row)

select * from dd_singlecol_bitmap_idx where (a=1 or a=2) and (b=2 or c=2);
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 1 0
a | b | c
---+---+---
2 | 2 | 2
Expand Down
2 changes: 1 addition & 1 deletion contrib/pax_storage/src/test/regress/expected/create_index_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2013,7 +2013,7 @@ WHERE unique1 IN (1,42,7)
ORDER BY unique1;
QUERY PLAN
-------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: unique1
-> Index Only Scan using tenk1_unique1 on tenk1
Index Cond: (unique1 = ANY ('{1,42,7}'::integer[]))
Expand Down
16 changes: 8 additions & 8 deletions contrib/pax_storage/src/test/regress/expected/direct_dispatch_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1120,14 +1120,14 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
explain (costs off) select gp_segment_id, * from t1_varchar where col1_varchar in ('a','b');
QUERY PLAN
----------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Seq Scan on t1_varchar
Filter: ((col1_varchar)::text = ANY ('{a,b}'::text[]))
Optimizer: Pivotal Optimizer (GPORCA)
(4 rows)

select gp_segment_id, * from t1_varchar where col1_varchar in ('a','b');
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
gp_segment_id | col1_varchar | col2_int
---------------+--------------+----------
1 | b | 2
Expand All @@ -1137,14 +1137,14 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
explain (costs off) select gp_segment_id, * from t1_varchar where col1_varchar = 'a' or col1_varchar = 'b';
QUERY PLAN
--------------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Seq Scan on t1_varchar
Filter: (((col1_varchar)::text = 'a'::text) OR ((col1_varchar)::text = 'b'::text))
Optimizer: Pivotal Optimizer (GPORCA)
(4 rows)

select gp_segment_id, * from t1_varchar where col1_varchar = 'a' or col1_varchar = 'b';
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
gp_segment_id | col1_varchar | col2_int
---------------+--------------+----------
1 | b | 2
Expand Down Expand Up @@ -1204,14 +1204,14 @@ INFO: (slice 1) Dispatch command to SINGLE content
explain (costs off) select gp_segment_id, * from t1_varchar where col1_varchar in ('a', 'b') and col2_int=2;
QUERY PLAN
-------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Seq Scan on t1_varchar
Filter: (((col1_varchar)::text = ANY ('{a,b}'::text[])) AND (col2_int = 2))
Optimizer: Pivotal Optimizer (GPORCA)
(4 rows)

select gp_segment_id, * from t1_varchar where col1_varchar in ('a', 'b') and col2_int=2;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
gp_segment_id | col1_varchar | col2_int
---------------+--------------+----------
1 | b | 2
Expand All @@ -1220,14 +1220,14 @@ INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
explain (costs off) select gp_segment_id, * from t1_varchar where (col1_varchar = 'a' or col1_varchar = 'b') and col2_int=1;
QUERY PLAN
-----------------------------------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Seq Scan on t1_varchar
Filter: ((((col1_varchar)::text = 'a'::text) OR ((col1_varchar)::text = 'b'::text)) AND (col2_int = 1))
Optimizer: Pivotal Optimizer (GPORCA)
(4 rows)

select gp_segment_id, * from t1_varchar where (col1_varchar = 'a' or col1_varchar = 'b') and col2_int=1;
INFO: (slice 1) Dispatch command to ALL contents: 0 1 2
INFO: (slice 1) Dispatch command to PARTIAL contents: 2 1
gp_segment_id | col1_varchar | col2_int
---------------+--------------+----------
2 | a | 1
Expand Down
16 changes: 8 additions & 8 deletions contrib/pax_storage/src/test/regress/expected/inherit_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2088,7 +2088,7 @@ explain (costs off) select * from list_parted where a is not null;
explain (costs off) select * from list_parted where a in ('ab', 'cd', 'ef');
QUERY PLAN
----------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on list_parted
Number of partitions to scan: 2 (out of 3)
Filter: ((a)::text = ANY ('{ab,cd,ef}'::text[]))
Expand All @@ -2098,7 +2098,7 @@ explain (costs off) select * from list_parted where a in ('ab', 'cd', 'ef');
explain (costs off) select * from list_parted where a = 'ab' or a in (null, 'cd');
QUERY PLAN
---------------------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on list_parted
Number of partitions to scan: 1 (out of 3)
Filter: (((a)::text = 'ab'::text) OR ((a)::text = ANY ('{NULL,cd}'::text[])))
Expand Down Expand Up @@ -2548,7 +2548,7 @@ explain (costs off) select * from mclparted order by a;
explain (costs off) select * from mclparted where a in(3,4,5) order by a;
QUERY PLAN
---------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand All @@ -2565,7 +2565,7 @@ create table mclparted_def partition of mclparted default;
explain (costs off) select * from mclparted where a in(1,2,4) order by a;
QUERY PLAN
---------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand All @@ -2578,7 +2578,7 @@ explain (costs off) select * from mclparted where a in(1,2,4) order by a;
explain (costs off) select * from mclparted where a in(1,2,4) or a is null order by a;
QUERY PLAN
-------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand All @@ -2595,7 +2595,7 @@ create table mclparted_0_null partition of mclparted for values in(0,null);
explain (costs off) select * from mclparted where a in(1,2,4) or a is null order by a;
QUERY PLAN
-------------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand All @@ -2608,7 +2608,7 @@ explain (costs off) select * from mclparted where a in(1,2,4) or a is null order
explain (costs off) select * from mclparted where a in(0,1,2,4) order by a;
QUERY PLAN
---------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand All @@ -2622,7 +2622,7 @@ explain (costs off) select * from mclparted where a in(0,1,2,4) order by a;
explain (costs off) select * from mclparted where a in(1,2,4) order by a;
QUERY PLAN
---------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
Merge Key: a
-> Sort
Sort Key: a
Expand Down
8 changes: 4 additions & 4 deletions contrib/pax_storage/src/test/regress/expected/partition_prune_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1600,7 +1600,7 @@ create table coercepart_cd partition of coercepart for values in ('cd');
explain (costs off) select * from coercepart where a in ('ab', to_char(125, '999'));
QUERY PLAN
-----------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on coercepart
Number of partitions to scan: 1 (out of 3)
Filter: ((a)::text = ANY ('{ab," 125"}'::text[]))
Expand Down Expand Up @@ -1650,7 +1650,7 @@ explain (costs off) select * from coercepart where a !~ all ('{ab,bc}');
explain (costs off) select * from coercepart where a = any ('{ab,bc}');
QUERY PLAN
-------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on coercepart
Number of partitions to scan: 2 (out of 3)
Filter: ((a)::text = ANY ('{ab,bc}'::text[]))
Expand Down Expand Up @@ -1690,7 +1690,7 @@ explain (costs off) select * from coercepart where a = all ('{ab}');
explain (costs off) select * from coercepart where a = all ('{ab,bc}');
QUERY PLAN
-------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on coercepart
Number of partitions to scan: 2 (out of 3)
Filter: ((a)::text = ALL ('{ab,bc}'::text[]))
Expand Down Expand Up @@ -3894,7 +3894,7 @@ explain (costs off) select * from pp_arrpart where a = '{1, 2}';
explain (costs off) select * from pp_arrpart where a in ('{4, 5}', '{1}');
QUERY PLAN
----------------------------------------------------------------------
Gather Motion 3:1 (slice1; segments: 3)
Gather Motion 2:1 (slice1; segments: 2)
-> Dynamic Seq Scan on pp_arrpart
Number of partitions to scan: 2 (out of 2)
Filter: ((a = '{4,5}'::integer[]) OR (a = '{1}'::integer[]))
Expand Down
2 changes: 1 addition & 1 deletion contrib/pax_storage/src/test/regress/expected/qp_misc_jiras_optimizer.out
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5635,7 +5635,7 @@ explain select b from epsilon_test where b in (11,30) limit 30;
QUERY PLAN
--------------------------------------------------------------------------------------
Limit (cost=0.00..8771.18 rows=10 width=2)
-> Gather Motion 3:1 (slice1; segments: 3) (cost=0.00..8771.18 rows=30 width=2)
-> Gather Motion 2:1 (slice1; segments: 2) (cost=0.00..8771.18 rows=30 width=2)
-> Limit (cost=0.00..8771.18 rows=10 width=2)
-> Seq Scan on epsilon_test (cost=0.00..8771.18 rows=10677 width=2)
Filter: (b = ANY ('{11,30}'::integer[]))
Expand Down
11 changes: 11 additions & 0 deletions src/backend/gpopt/gpdbwrappers.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1204,6 +1204,17 @@ gpdb::LAppendInt(List *list, int iDatum)
return NIL;
}

List *
gpdb::LAppendUniqueInt(List *list, int iDatum)
{
GP_WRAP_START;
{
return list_append_unique_int(list, iDatum);
}
GP_WRAP_END;
return NIL;
}

List *
gpdb::LAppendOid(List *list, Oid datum)
{
Expand Down
27 changes: 15 additions & 12 deletions src/backend/gpopt/translate/CTranslatorDXLToPlStmt.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5530,9 +5530,6 @@ CTranslatorDXLToPlStmt::TranslateDXLDirectDispatchInfo(
return NIL;
}

CDXLDatumArray *dxl_datum_array = (*dispatch_identifier_datum_arrays)[0];
GPOS_ASSERT(0 < dxl_datum_array->Size());

const ULONG length = dispatch_identifier_datum_arrays->Size();

if (dxl_direct_dispatch_info->FContainsRawValues())
Expand Down Expand Up @@ -5568,23 +5565,29 @@ CTranslatorDXLToPlStmt::TranslateDXLDirectDispatchInfo(
return segids_list;
}

ULONG hash_code = GetDXLDatumGPDBHash(dxl_datum_array, pRTEHashFuncCal);
// Each datum array is one value combination of the distribution key
// (e.g. one array per IN-list constant). Dispatch to the union of the
// segments they hash to, like the planner does.
List *segids_list = NIL;
for (ULONG ul = 0; ul < length; ul++)
{
CDXLDatumArray *dispatch_identifier_datum_array =
(*dispatch_identifier_datum_arrays)[ul];
GPOS_ASSERT(0 < dispatch_identifier_datum_array->Size());
ULONG hash_code_new = GetDXLDatumGPDBHash(
dispatch_identifier_datum_array, pRTEHashFuncCal);
ULONG hash_code = GetDXLDatumGPDBHash(dispatch_identifier_datum_array,
pRTEHashFuncCal);

if (hash_code != hash_code_new)
{
// values don't hash to the same segment
return NIL;
}
segids_list = gpdb::LAppendUniqueInt(segids_list, hash_code);
}

// a value set that spans every segment is equivalent to not direct
// dispatching at all
if (gpdb::ListLength(segids_list) == m_num_of_segments)
{
gpdb::ListFree(segids_list);
return NIL;
}

List *segids_list = gpdb::LAppendInt(NIL, hash_code);
return segids_list;
}

Expand Down
3 changes: 3 additions & 0 deletions src/include/gpopt/gpdbwrappers.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -456,6 +456,9 @@ List *LAppend(List *list, void *datum);
// append an integer to a list
List *LAppendInt(List *list, int datum);

// append an integer to a list if it is not already a member
List *LAppendUniqueInt(List *list, int datum);

// append an oid to a list
List *LAppendOid(List *list, Oid datum);

Expand Down
Loading
Loading