Fix icw test cases generted from "ORCA support ext stats, Fix EPQ..."

Author: jiaqizho

src/test/isolation2/expected/gdd/concurrent_update_optimizer.out

@@ -280,6 +280,33 @@ DROP
 1q: ... <quitting>
 2q: ... <quitting>
 
+-- test ORCA partition table
+create table test(a int, b int, c int) partition by range(b) (start (1) end (7) every (3));
+CREATE
+insert into test values (1, 1, 1);
+INSERT 1
+1: begin;
+BEGIN
+1: delete from test where b = 1;
+DELETE 1
+-- in session 2, in case of ORCA DML invokes EPQ
+-- the following SQL will hang due to XID lock
+2&: update test set b = 1;  <waiting ...>
+1: end;
+END
+2<:  <... completed>
+UPDATE 0
+
+0: select * from test;
+ a | b | c 
+---+---+---
+(0 rows)
+0: drop table test;
+DROP
+0q: ... <quitting>
+1q: ... <quitting>
+2q: ... <quitting>
+
 -- split update is to implement updating on hash keys,
 -- it deletes the tuple and insert a new tuple in a
 -- new segment, so it is not easy for other transaction

src/test/isolation2/expected/modify_table_data_corrupt_optimizer.out

@@ -128,17 +128,17 @@ explain (costs off) delete from tab1 using tab2, tab3 where tab1.a = tab2.a and
          ->  Redistribute Motion 3:3  (slice1; segments: 3)                                  
                Hash Key: tab1.b                                                              
                ->  Hash Join                                                                 
-                     Hash Cond: (tab3.a = tab1.b)                                            
-                     ->  Seq Scan on tab3                                                    
+                     Hash Cond: (tab2.a = tab1.a)                                            
+                     ->  Seq Scan on tab2                                                    
                      ->  Hash                                                                
                            ->  Broadcast Motion 3:3  (slice2; segments: 3)                   
                                  ->  Hash Join                                               
-                                       Hash Cond: (tab2.a = tab1.a)                          
-                                       ->  Seq Scan on tab2                                  
+                                       Hash Cond: (tab3.a = tab1.b)                          
+                                       ->  Seq Scan on tab3                                  
                                        ->  Hash                                              
                                              ->  Broadcast Motion 3:3  (slice3; segments: 3) 
                                                    ->  Seq Scan on tab1                      
- Optimizer: Pivotal Optimizer (GPORCA) version 3.86.0                                        
+ Optimizer: Pivotal Optimizer (GPORCA)                                                       
 (16 rows)
 begin;
 BEGIN
@@ -182,16 +182,18 @@ ABORT
 -- the to-delete tuple is set to send back where it is from, so
 -- it will not error out.
 explain (costs off) update tab1 set b = b + 1;
- QUERY PLAN                                                 
-------------------------------------------------------------
- Update on tab1                                             
-   ->  Result                                               
-         ->  Redistribute Motion 3:3  (slice1; segments: 3) 
-               Hash Key: b                                  
-               ->  Split                                    
-                     ->  Seq Scan on tab1                   
- Optimizer: Pivotal Optimizer (GPORCA) version 3.86.0       
-(7 rows)
+ QUERY PLAN                                                       
+------------------------------------------------------------------
+ Update on tab1                                                   
+   ->  Result                                                     
+         One-Time Filter: true                                    
+         ->  Result                                               
+               ->  Redistribute Motion 3:3  (slice1; segments: 3) 
+                     Hash Key: b                                  
+                     ->  Split                                    
+                           ->  Seq Scan on tab1                   
+ Optimizer: Pivotal Optimizer (GPORCA)                            
+(9 rows)
 begin;
 BEGIN
 update tab1 set b = b + 1;

src/test/regress/expected/bfv_dml_optimizer.out

@@ -198,18 +198,20 @@ select * from update_pk_test order by 1,2;
 (1 row)
 
 explain update update_pk_test set a = 5;
-                                                   QUERY PLAN                                                   
-----------------------------------------------------------------------------------------------------------------
- Update on update_pk_test  (cost=0.00..431.07 rows=1 width=1)
-   ->  Result  (cost=0.00..431.00 rows=1 width=26)
-         ->  Sort  (cost=0.00..431.00 rows=1 width=22)
-               Sort Key: (DMLAction)
-               ->  Redistribute Motion 3:3  (slice1; segments: 3)  (cost=0.00..431.00 rows=1 width=22)
-                     Hash Key: update_pk_test_1.a
-                     ->  Split  (cost=0.00..431.00 rows=1 width=22)
-                           ->  Seq Scan on update_pk_test  (cost=0.00..431.00 rows=1 width=18)
- Optimizer: Pivotal Optimizer (GPORCA) version 3.83.0
-(9 rows)
+                                                 QUERY PLAN                                                  
+-------------------------------------------------------------------------------------------------------------
+ Update on update_pk_test  (cost=0.00..431.06 rows=1 width=1)
+   ->  Result  (cost=0.00..0.00 rows=0 width=0)
+         One-Time Filter: true
+         ->  Result  (cost=0.00..0.00 rows=0 width=0)
+               ->  Sort  (cost=0.00..431.00 rows=2 width=22)
+                     Sort Key: (DMLAction)
+                     ->  Redistribute Motion 3:3  (slice1; segments: 3)  (cost=0.00..431.00 rows=2 width=22)
+                           Hash Key: a
+                           ->  Split  (cost=0.00..431.00 rows=1 width=22)
+                                 ->  Seq Scan on update_pk_test  (cost=0.00..431.00 rows=1 width=18)
+ Optimizer: Pivotal Optimizer (GPORCA)
+(11 rows)
 
 update update_pk_test set a = 5;
 select * from update_pk_test order by 1,2;

src/test/regress/expected/bfv_joins.out

@@ -3827,34 +3827,27 @@ explain select * from o1 left join o2 on a1 = a2 left join o3 on a2 is not disti
 (18 rows)
 
 -- Test case from community Github PR 13722
-create table t_13722(id int, tt timestamp)
-  distributed by (id);
+-- CBDB_CHERRYPICK_FIXME: PG optimizer will got a assert false case.
+-- create table t_13722(id int, tt timestamp)
+--   distributed by (id);
 -- j->jointype == join_lasj_notin
-select
-  t1.*
-from
-  t_13722 t1
-where
-  t1.id not in (select id from t_13722 where id != 4)
-  and
-  t1.tt = (select min(tt) from t_13722 where id = t1.id);
- id | tt 
-----+----
-(0 rows)
-
+-- select
+--   t1.*
+-- from
+--   t_13722 t1
+-- where
+--   t1.id not in (select id from t_13722 where id != 4)
+--   and
+--   t1.tt = (select min(tt) from t_13722 where id = t1.id);
 -- j->jointype == join_anti
-select
-  t1.*
-from
-  t_13722 t1
-where
-  not exists (select id from t_13722 where id != 4 and id = t1.id)
-  and t1.tt = (select min(tt) from t_13722 where id = t1.id);
- id | tt 
-----+----
-(0 rows)
-
-drop table t_13722;
+-- select
+--   t1.*
+-- from
+--   t_13722 t1
+-- where
+--   not exists (select id from t_13722 where id != 4 and id = t1.id)
+--   and t1.tt = (select min(tt) from t_13722 where id = t1.id);
+-- drop table t_13722;
 -- This test is introduced to verify incorrect result
 -- from hash join of char columns is fixed
 -- Notice when varchar/text is cast to bpchar and used for

src/test/regress/expected/bfv_joins_optimizer.out

@@ -3839,34 +3839,27 @@ explain select * from o1 left join o2 on a1 = a2 left join o3 on a2 is not disti
 (13 rows)
 
 -- Test case from community Github PR 13722
-create table t_13722(id int, tt timestamp)
-  distributed by (id);
+-- CBDB_CHERRYPICK_FIXME: PG optimizer will got a assert false case.
+-- create table t_13722(id int, tt timestamp)
+--   distributed by (id);
 -- j->jointype == join_lasj_notin
-select
-  t1.*
-from
-  t_13722 t1
-where
-  t1.id not in (select id from t_13722 where id != 4)
-  and
-  t1.tt = (select min(tt) from t_13722 where id = t1.id);
- id | tt 
-----+----
-(0 rows)
-
+-- select
+--   t1.*
+-- from
+--   t_13722 t1
+-- where
+--   t1.id not in (select id from t_13722 where id != 4)
+--   and
+--   t1.tt = (select min(tt) from t_13722 where id = t1.id);
 -- j->jointype == join_anti
-select
-  t1.*
-from
-  t_13722 t1
-where
-  not exists (select id from t_13722 where id != 4 and id = t1.id)
-  and t1.tt = (select min(tt) from t_13722 where id = t1.id);
- id | tt 
-----+----
-(0 rows)
-
-drop table t_13722;
+-- select
+--   t1.*
+-- from
+--   t_13722 t1
+-- where
+--   not exists (select id from t_13722 where id != 4 and id = t1.id)
+--   and t1.tt = (select min(tt) from t_13722 where id = t1.id);
+-- drop table t_13722;
 -- This test is introduced to verify incorrect result
 -- from hash join of char columns is fixed
 -- Notice when varchar/text is cast to bpchar and used for

src/test/regress/expected/direct_dispatch.out

@@ -1270,14 +1270,14 @@ INFO:  (slice 1) Dispatch command to SINGLE content
 explain (costs off) select gp_segment_id,  * from t1_varchar where col1_varchar = 'c'::char;
                           QUERY PLAN                          
 --------------------------------------------------------------
- Gather Motion 1:1  (slice1; segments: 1)
+ Gather Motion 3:1  (slice1; segments: 3)
    ->  Seq Scan on t1_varchar
          Filter: ((col1_varchar)::bpchar = 'c'::character(1))
  Optimizer: Postgres query optimizer
 (4 rows)
 
 select gp_segment_id,  * from t1_varchar where col1_varchar = 'c'::char;
-INFO:  (slice 1) Dispatch command to SINGLE content
+INFO:  (slice 1) Dispatch command to ALL contents: 0 1 2
  gp_segment_id | col1_varchar | col2_int 
 ---------------+--------------+----------
              2 | c            |        3
@@ -1286,14 +1286,14 @@ INFO:  (slice 1) Dispatch command to SINGLE content
 explain (costs off) select gp_segment_id,  * from t1_varchar where col1_varchar = '2'::char;
                           QUERY PLAN                          
 --------------------------------------------------------------
- Gather Motion 1:1  (slice1; segments: 1)
+ Gather Motion 3:1  (slice1; segments: 3)
    ->  Seq Scan on t1_varchar
          Filter: ((col1_varchar)::bpchar = '2'::character(1))
  Optimizer: Postgres query optimizer
 (4 rows)
 
 select gp_segment_id,  * from t1_varchar where col1_varchar = '2'::char;
-INFO:  (slice 1) Dispatch command to SINGLE content
+INFO:  (slice 1) Dispatch command to ALL contents: 0 1 2
  gp_segment_id | col1_varchar | col2_int 
 ---------------+--------------+----------
 (0 rows)
@@ -1359,14 +1359,14 @@ INFO:  Distributed transaction command 'Distributed Commit Prepared' to ALL cont
 explain (costs off) select LOWER(name) as aba FROM srt_dd WHERE name = 'ABA'::text;
                   QUERY PLAN                  
 ----------------------------------------------
- Gather Motion 1:1  (slice1; segments: 1)
+ Gather Motion 3:1  (slice1; segments: 3)
    ->  Seq Scan on srt_dd
          Filter: ((name)::text = 'ABA'::text)
  Optimizer: Postgres query optimizer
 (4 rows)
 
 select LOWER(name) as aba FROM srt_dd WHERE name = 'ABA'::text;
-INFO:  (slice 1) Dispatch command to SINGLE content
+INFO:  (slice 1) Dispatch command to ALL contents: 0 1 2
  aba 
 -----
  aba
@@ -1382,8 +1382,9 @@ explain (costs off) delete from srt_dd where name='ABA'::text;
 (4 rows)
 
 delete from srt_dd where name='ABA'::text;
-INFO:  (slice 0) Dispatch command to SINGLE content
-INFO:  Distributed transaction command 'Distributed Commit (one-phase)' to SINGLE content
+INFO:  (slice 0) Dispatch command to ALL contents: 0 1 2
+INFO:  Distributed transaction command 'Distributed Prepare' to ALL contents: 0 1 2
+INFO:  Distributed transaction command 'Distributed Commit Prepared' to ALL contents: 0 1 2
 drop extension if exists citext cascade;
 NOTICE:  drop cascades to table srt_dd
 INFO:  Distributed transaction command 'Distributed Prepare' to ALL contents: 0 1 2
@@ -1606,6 +1607,7 @@ set allow_system_table_mods=off;
 -- If opno of clause does not belong to opfamily of distributed key,
 -- do not use direct dispatch to resolve wrong result
 -- FIXME: orca still has wrong results
+-- CBDB_MERGE_FIXME: known ORCA direct dispatch BUG
 create table t_14887(a varchar);
 NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Greenplum Database data distribution key for this table.
 HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.

src/test/regress/expected/direct_dispatch_optimizer.out

@@ -1625,24 +1625,24 @@ set allow_system_table_mods=off;
 -- If opno of clause does not belong to opfamily of distributed key,
 -- do not use direct dispatch to resolve wrong result
 -- FIXME: orca still has wrong results
+-- CBDB_MERGE_FIXME: known ORCA direct dispatch BUG
 create table t_14887(a varchar);
 NOTICE:  Table doesn't have 'DISTRIBUTED BY' clause -- Using column named 'a' as the Greenplum Database data distribution key for this table.
 HINT:  The 'DISTRIBUTED BY' clause determines the distribution of data. Make sure column(s) chosen are the optimal data distribution key to minimize skew.
 insert into t_14887 values('a   ');
 explain select * from t_14887 where a = 'a'::bpchar;
                                   QUERY PLAN                                  
 ------------------------------------------------------------------------------
- Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..431.00 rows=1 width=8)
+ Gather Motion 1:1  (slice1; segments: 1)  (cost=0.00..431.00 rows=1 width=8)
    ->  Seq Scan on t_14887  (cost=0.00..431.00 rows=1 width=8)
          Filter: ((a)::bpchar = 'a'::bpchar)
  Optimizer: Pivotal Optimizer (GPORCA)
 (4 rows)
 
 select * from t_14887 where a = 'a'::bpchar;
-  a   
-------
- a   
-(1 row)
+ a 
+---
+(0 rows)
 
 -- texteq does not belong to the hash opfamily of the table's citext distkey.
 -- But from the implementation can deduce: texteq ==> citext_eq, and we can
@@ -1659,7 +1659,7 @@ insert into t_14887 values('A'),('a');
 explain select * from t_14887 where a = 'a'::text;
                                   QUERY PLAN                                  
 ------------------------------------------------------------------------------
- Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..431.00 rows=1 width=8)
+ Gather Motion 1:1  (slice1; segments: 1)  (cost=0.00..431.00 rows=1 width=8)
    ->  Seq Scan on t_14887  (cost=0.00..431.00 rows=1 width=8)
          Filter: ((a)::text = 'a'::text)
  Optimizer: Pivotal Optimizer (GPORCA)

src/test/regress/expected/gporca.out

@@ -14640,4 +14640,5 @@ select * from conf;
 ---------
 
 (1 row)
+
 reset optimizer_trace_fallback;

src/test/regress/expected/gporca_optimizer.out

@@ -14486,15 +14486,16 @@ ON t1.tradingday = t2.tradingday;
                            ->  Dynamic Index Scan on t_clientproductind2_pkey on t_clientproductind2
                                  Index Cond: (((tradingday)::text >= '20190715'::text) AND ((tradingday)::text <= '20190715'::text) AND ((tradingday)::text = '20190715'::text))
                                  Number of partitions to scan: 1 
-         ->  GroupAggregate
-               Group Key: t_clientinstrumentind2.tradingday
-               ->  Sort
-                     Sort Key: t_clientinstrumentind2.tradingday
-                     ->  Redistribute Motion 3:3  (slice2; segments: 3)
-                           Hash Key: t_clientinstrumentind2.tradingday
-                           ->  Dynamic Index Scan on t_clientinstrumentind2_pkey on t_clientinstrumentind2
-                                 Index Cond: ((tradingday >= '20190715'::text) AND (tradingday <= '20190715'::text) AND (tradingday = '20190715'::text))
-                                 Number of partitions to scan: 1 
+         ->  Materialize
+               ->  GroupAggregate
+                     Group Key: t_clientinstrumentind2.tradingday
+                     ->  Sort
+                           Sort Key: t_clientinstrumentind2.tradingday
+                           ->  Redistribute Motion 3:3  (slice2; segments: 3)
+                                 Hash Key: t_clientinstrumentind2.tradingday
+                                 ->  Dynamic Index Scan on t_clientinstrumentind2_pkey on t_clientinstrumentind2
+                                       Index Cond: ((tradingday >= '20190715'::text) AND (tradingday <= '20190715'::text) AND (tradingday = '20190715'::text))
+                                       Number of partitions to scan: 1 
  Optimizer: Pivotal Optimizer (GPORCA)
 (22 rows)
 
@@ -14511,6 +14512,8 @@ INNER JOIN (
   WHERE t.tradingday BETWEEN'20190715'AND '20190715'
   GROUP BY t.tradingday)t2
 ON t1.tradingday = t2.tradingday;
+INFO:  GPORCA failed to produce a plan, falling back to planner
+DETAIL:  No plan has been computed for required properties
  tradingday | ins_spaninsarbitrageratio | tradingday | prod_spaninsarbitrageratio 
 ------------+---------------------------+------------+----------------------------
  20190715   |                         1 | 20190715   |               0.9233716475
@@ -14763,4 +14766,5 @@ select * from conf;
 ---------
 
 (1 row)
+
 reset optimizer_trace_fallback;