Consider predicate of type <cast(ident)> array cmp <const array>

While processing constraint interval, also consider predicate of type
<cast(ident)> array cmp else we lose the opportunity
to generate implied quals. This is corresponding to the ORCA changes.

concourse/tasks/compile_gpdb.yml

@@ -19,4 +19,4 @@ params:
   BLD_TARGETS:
   OUTPUT_ARTIFACT_DIR: gpdb_artifacts
   CONFIGURE_FLAGS:
-  ORCA_TAG: v3.66.0
+  ORCA_TAG: v3.67.0

config/orca.m4

@@ -40,10 +40,10 @@ AC_RUN_IFELSE([AC_LANG_PROGRAM([[
 #include <string.h>
 ]],
 [
-return strncmp("3.66.", GPORCA_VERSION_STRING, 5);
+return strncmp("3.67.", GPORCA_VERSION_STRING, 5);
 ])],
 [AC_MSG_RESULT([[ok]])],
-[AC_MSG_ERROR([Your ORCA version is expected to be 3.66.XXX])]
+[AC_MSG_ERROR([Your ORCA version is expected to be 3.67.XXX])]
 )
 AC_LANG_POP([C++])
 ])# PGAC_CHECK_ORCA_VERSION

configure

@@ -14902,7 +14902,7 @@ int
 main ()
 {
 
-return strncmp("3.66.", GPORCA_VERSION_STRING, 5);
+return strncmp("3.67.", GPORCA_VERSION_STRING, 5);
 
   ;
   return 0;
@@ -14912,7 +14912,7 @@ if ac_fn_cxx_try_run "$LINENO"; then :
   { $as_echo "$as_me:${as_lineno-$LINENO}: result: ok" >&5
 $as_echo "ok" >&6; }
 else
-  as_fn_error $? "Your ORCA version is expected to be 3.66.XXX" "$LINENO" 5
+  as_fn_error $? "Your ORCA version is expected to be 3.67.XXX" "$LINENO" 5
 
 fi
 rm -f core *.core core.conftest.* gmon.out bb.out conftest$ac_exeext \

depends/conanfile_orca.txt

 [requires]
-orca/v3.66.0@gpdb/stable
+orca/v3.67.0@gpdb/stable
 
 [imports]
 include, * -> build/include

src/test/regress/expected/gporca.out

@@ -11383,3 +11383,53 @@ SELECT *, a IN ( SELECT b + 1 FROM non_part1) FROM ds_part, non_part2 WHERE ds_p
 (1 row)
 
 RESET optimizer_enforce_subplans;
+-- implied predicate must be generated for the type cast(ident) scalar array cmp const array
+CREATE TABLE varchar_sc_array_cmp(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 varchar_sc_array_cmp VALUES ('a'), ('b'), ('c'), ('d');
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+                                       QUERY PLAN                                        
+-----------------------------------------------------------------------------------------
+ Gather Motion 2:1  (slice1; segments: 2)  (cost=3.07..6.17 rows=4 width=4)
+   ->  Hash Join  (cost=3.07..6.17 rows=2 width=4)
+         Hash Cond: ((t1.a)::text = (t2.a)::text)
+         ->  Seq Scan on varchar_sc_array_cmp t1  (cost=0.00..3.05 rows=1 width=2)
+               Filter: ((a)::text = ANY ('{b,c}'::text[]))
+         ->  Hash  (cost=3.05..3.05 rows=1 width=2)
+               ->  Seq Scan on varchar_sc_array_cmp t2  (cost=0.00..3.05 rows=1 width=2)
+                     Filter: ((a)::text = ANY ('{b,c}'::text[]))
+ Optimizer: Postgres query optimizer
+(9 rows)
+
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+ a | a 
+---+---
+ c | c
+ b | b
+(2 rows)
+
+SET optimizer_array_constraints=on;
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+                                          QUERY PLAN                                          
+----------------------------------------------------------------------------------------------
+ Gather Motion 3:1  (slice1; segments: 3)  (cost=3.09..6.18 rows=4 width=4)
+   ->  Hash Join  (cost=3.09..6.18 rows=2 width=4)
+         Hash Cond: ((t2.a)::text = (t1.a)::text)
+         ->  Seq Scan on varchar_sc_array_cmp t2  (cost=0.00..3.04 rows=2 width=2)
+         ->  Hash  (cost=3.06..3.06 rows=1 width=2)
+               ->  Seq Scan on varchar_sc_array_cmp t1  (cost=0.00..3.06 rows=1 width=2)
+                     Filter: (((a)::text = ANY ('{b,c}'::text[])) OR ((a)::text = 'a'::text))
+ Optimizer: Postgres query optimizer
+(8 rows)
+
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+ a | a 
+---+---
+ b | b
+ a | a
+ c | c
+(3 rows)
+
+DROP TABLE varchar_sc_array_cmp;
+--

src/test/regress/expected/gporca_optimizer.out

@@ -11480,6 +11480,57 @@ SELECT *, a IN ( SELECT b + 1 FROM non_part1) FROM ds_part, non_part2 WHERE ds_p
 (1 row)
 
 RESET optimizer_enforce_subplans;
+-- implied predicate must be generated for the type cast(ident) scalar array cmp const array
+CREATE TABLE varchar_sc_array_cmp(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 varchar_sc_array_cmp VALUES ('a'), ('b'), ('c'), ('d');
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+                                                      QUERY PLAN                                                      
+----------------------------------------------------------------------------------------------------------------------
+ Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..862.00 rows=4 width=4)
+   ->  Hash Join  (cost=0.00..862.00 rows=2 width=4)
+         Hash Cond: ((varchar_sc_array_cmp.a)::text = (varchar_sc_array_cmp_1.a)::text)
+         ->  Seq Scan on varchar_sc_array_cmp  (cost=0.00..431.00 rows=1 width=2)
+               Filter: (((a)::text = ANY ('{b,c}'::text[])) AND (((a)::text = 'b'::text) OR ((a)::text = 'c'::text)))
+         ->  Hash  (cost=431.00..431.00 rows=1 width=2)
+               ->  Seq Scan on varchar_sc_array_cmp varchar_sc_array_cmp_1  (cost=0.00..431.00 rows=1 width=2)
+                     Filter: (((a)::text = 'b'::text) OR ((a)::text = 'c'::text))
+ Optimizer: Pivotal Optimizer (GPORCA) version 3.65.2
+(9 rows)
+
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+ a | a 
+---+---
+ b | b
+ c | c
+(2 rows)
+
+SET optimizer_array_constraints=on;
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+                                                  QUERY PLAN                                                   
+---------------------------------------------------------------------------------------------------------------
+ Gather Motion 3:1  (slice1; segments: 3)  (cost=0.00..862.00 rows=4 width=4)
+   ->  Hash Join  (cost=0.00..862.00 rows=2 width=4)
+         Hash Cond: ((varchar_sc_array_cmp.a)::text = (varchar_sc_array_cmp_1.a)::text)
+         ->  Seq Scan on varchar_sc_array_cmp  (cost=0.00..431.00 rows=1 width=2)
+               Filter: (a = ANY ('{a,b,c}'::character varying[]))
+         ->  Hash  (cost=431.00..431.00 rows=1 width=2)
+               ->  Seq Scan on varchar_sc_array_cmp varchar_sc_array_cmp_1  (cost=0.00..431.00 rows=1 width=2)
+                     Filter: (a = ANY ('{a,b,c}'::character varying[]))
+ Optimizer: Pivotal Optimizer (GPORCA) version 3.65.2
+(9 rows)
+
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+ a | a 
+---+---
+ a | a
+ c | c
+ b | b
+(3 rows)
+
+DROP TABLE varchar_sc_array_cmp;
+-- 
 -- start_ignore
 DROP SCHEMA orca CASCADE;
 NOTICE:  drop cascades to 92 other objects

src/test/regress/sql/gporca.sql

@@ -2184,6 +2184,16 @@ analyze ds_part;
 SELECT *, a IN ( SELECT b + 1 FROM non_part1) FROM ds_part, non_part2 WHERE ds_part.c = non_part2.e AND non_part2.f = 10 AND a IN ( SELECT b FROM non_part1);
 
 RESET optimizer_enforce_subplans;
+-- implied predicate must be generated for the type cast(ident) scalar array cmp const array
+CREATE TABLE varchar_sc_array_cmp(a varchar);
+INSERT INTO varchar_sc_array_cmp VALUES ('a'), ('b'), ('c'), ('d');
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and t1.a in ('b', 'c');
+SET optimizer_array_constraints=on;
+EXPLAIN SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+SELECT * FROM varchar_sc_array_cmp t1, varchar_sc_array_cmp t2 where t1.a = t2.a and (t1.a in ('b', 'c') OR t1.a = 'a');
+DROP TABLE varchar_sc_array_cmp;
+-- 
 
 -- start_ignore
 DROP SCHEMA orca CASCADE;