optbuilder: try harder to propagate types in UNION

Release note: None

pkg/sql/logictest/testdata/logic_test/union

@@ -209,21 +209,24 @@ SELECT 1, 2 INTERSECT SELECT 3
 query error pgcode 42601 each EXCEPT query must have the same number of columns: 2 vs 1
 SELECT 1, 2 EXCEPT SELECT 3
 
-query error pgcode 42804 UNION types int and string cannot be matched
-SELECT 1 UNION SELECT '3'
-
-query error pgcode 42804 INTERSECT types int and string cannot be matched
-SELECT 1 INTERSECT SELECT '3'
-
-query error pgcode 42804 EXCEPT types int and string cannot be matched
-SELECT 1 EXCEPT SELECT '3'
+# These work with the optimizer on, but not with it off. Skip for now.
+# TODO(jordan,radu): re-enable when optimizer=off goes away.
+
+# query error pgcode 42804 UNION types int and string cannot be matched
+# SELECT 1 UNION SELECT '3'
+#
+# query error pgcode 42804 INTERSECT types int and string cannot be matched
+# SELECT 1 INTERSECT SELECT '3'
+#
+# query error pgcode 42804 EXCEPT types int and string cannot be matched
+# SELECT 1 EXCEPT SELECT '3'
+# 
+# query error UNION types int\[] and string\[] cannot be matched
+# SELECT ARRAY[1] UNION ALL SELECT ARRAY['foo']
 
 query error pgcode 42703 column \"z\" does not exist
 SELECT 1 UNION SELECT 3 ORDER BY z
 
-query error UNION types int\[] and string\[] cannot be matched
-SELECT ARRAY[1] UNION ALL SELECT ARRAY['foo']
-
 # Check that UNION permits columns of different visible types
 
 statement ok

pkg/sql/opt/memo/testdata/stats/set

@@ -847,7 +847,7 @@ WHERE a IS NULL and b
 except
  ├── columns: a:5(int) b:2(bool!null)
  ├── left columns: column1:5(int) column2:2(bool!null)
- ├── right columns: column1:3(int) column2:6(bool)
+ ├── right columns: column1:3(int) column2:4(bool)
  ├── cardinality: [1 - 1]
  ├── stats: [rows=1, distinct(2,5)=1, null(2,5)=1]
  ├── key: (2,5)
@@ -859,11 +859,11 @@ except
  │    ├── fd: ()-->(2,5)
  │    └── (true, NULL) [type=tuple{bool, int}]
  └── values
-      ├── columns: column1:3(int!null) column2:6(bool!null)
+      ├── columns: column1:3(int!null) column2:4(bool!null)
       ├── cardinality: [0 - 0]
-      ├── stats: [rows=0, distinct(3,6)=0, null(3,6)=0]
+      ├── stats: [rows=0, distinct(3,4)=0, null(3,4)=0]
       ├── key: ()
-      └── fd: ()-->(3,6)
+      └── fd: ()-->(3,4)
 
 # Regression test for #36147.
 opt
@@ -920,7 +920,7 @@ sort
  └── intersect
       ├── columns: column1:1(int)
       ├── left columns: column1:1(int)
-      ├── right columns: column1:3(int)
+      ├── right columns: column1:2(int)
       ├── cardinality: [0 - 1]
       ├── stats: [rows=1, distinct(1)=1, null(1)=0]
       ├── key: (1)
@@ -932,9 +932,9 @@ sort
       │    ├── fd: ()-->(1)
       │    └── (1,) [type=tuple{int}]
       └── values
-           ├── columns: column1:3(int)
+           ├── columns: column1:2(int)
            ├── cardinality: [1 - 1]
-           ├── stats: [rows=1, distinct(3)=1, null(3)=1]
+           ├── stats: [rows=1, distinct(2)=1, null(2)=1]
            ├── key: ()
-           ├── fd: ()-->(3)
+           ├── fd: ()-->(2)
            └── (NULL,) [type=tuple{int}]

pkg/sql/opt/optbuilder/testdata/union

@@ -354,11 +354,11 @@ build
 SELECT x, pg_typeof(y) FROM (SELECT 1 AS a, 3 AS b UNION ALL SELECT 2 AS a, NULL AS b) AS t(x, y)
 ----
 project
- ├── columns: x:5(int!null) pg_typeof:8(string)
+ ├── columns: x:5(int!null) pg_typeof:7(string)
  ├── union-all
  │    ├── columns: a:5(int!null) b:6(int)
  │    ├── left columns: a:1(int) b:2(int)
- │    ├── right columns: a:3(int) b:7(int)
+ │    ├── right columns: a:3(int) b:4(int)
  │    ├── project
  │    │    ├── columns: a:1(int!null) b:2(int!null)
  │    │    ├── values
@@ -367,17 +367,13 @@ project
  │    │         ├── const: 1 [type=int]
  │    │         └── const: 3 [type=int]
  │    └── project
- │         ├── columns: b:7(int) a:3(int!null)
- │         ├── project
- │         │    ├── columns: a:3(int!null) b:4(unknown)
- │         │    ├── values
- │         │    │    └── tuple [type=tuple]
- │         │    └── projections
- │         │         ├── const: 2 [type=int]
- │         │         └── null [type=unknown]
+ │         ├── columns: a:3(int!null) b:4(int)
+ │         ├── values
+ │         │    └── tuple [type=tuple]
  │         └── projections
+ │              ├── const: 2 [type=int]
  │              └── cast: INT8 [type=int]
- │                   └── variable: b [type=unknown]
+ │                   └── null [type=unknown]
  └── projections
       └── function: pg_typeof [type=string]
            └── variable: b [type=int]
@@ -732,17 +728,62 @@ error (42601): each EXCEPT query must have the same number of columns: 2 vs 1
 build
 SELECT 1 UNION SELECT '3'
 ----
-error (42804): UNION types int and string cannot be matched
+union
+ ├── columns: "?column?":3(int!null)
+ ├── left columns: "?column?":1(int)
+ ├── right columns: "?column?":2(int)
+ ├── project
+ │    ├── columns: "?column?":1(int!null)
+ │    ├── values
+ │    │    └── tuple [type=tuple]
+ │    └── projections
+ │         └── const: 1 [type=int]
+ └── project
+      ├── columns: "?column?":2(int!null)
+      ├── values
+      │    └── tuple [type=tuple]
+      └── projections
+           └── const: 3 [type=int]
 
 build
 SELECT 1 INTERSECT SELECT '3'
 ----
-error (42804): INTERSECT types int and string cannot be matched
+intersect
+ ├── columns: "?column?":1(int!null)
+ ├── left columns: "?column?":1(int!null)
+ ├── right columns: "?column?":2(int)
+ ├── project
+ │    ├── columns: "?column?":1(int!null)
+ │    ├── values
+ │    │    └── tuple [type=tuple]
+ │    └── projections
+ │         └── const: 1 [type=int]
+ └── project
+      ├── columns: "?column?":2(int!null)
+      ├── values
+      │    └── tuple [type=tuple]
+      └── projections
+           └── const: 3 [type=int]
 
 build
 SELECT 1 EXCEPT SELECT '3'
 ----
-error (42804): EXCEPT types int and string cannot be matched
+except
+ ├── columns: "?column?":1(int!null)
+ ├── left columns: "?column?":1(int!null)
+ ├── right columns: "?column?":2(int)
+ ├── project
+ │    ├── columns: "?column?":1(int!null)
+ │    ├── values
+ │    │    └── tuple [type=tuple]
+ │    └── projections
+ │         └── const: 1 [type=int]
+ └── project
+      ├── columns: "?column?":2(int!null)
+      ├── values
+      │    └── tuple [type=tuple]
+      └── projections
+           └── const: 3 [type=int]
 
 build
 SELECT 1 UNION SELECT 3 ORDER BY z
@@ -752,7 +793,17 @@ error (42703): column "z" does not exist
 build
 SELECT ARRAY[1] UNION ALL SELECT ARRAY['foo']
 ----
-error (42804): UNION types int[] and string[] cannot be matched
+error (22P02): could not parse "foo" as type int: strconv.ParseInt: parsing "foo": invalid syntax
+
+build
+SELECT ARRAY['foo'] UNION ALL SELECT ARRAY[1]
+----
+error (42804): UNION types string[] and int[] cannot be matched
+
+build
+SELECT ARRAY[1] UNION ALL SELECT ARRAY[1.2]
+----
+error (42804): UNION types int[] and decimal[] cannot be matched
 
 exec-ddl
 CREATE TABLE t.xy (x STRING NOT NULL, y STRING NOT NULL)
@@ -874,7 +925,7 @@ VALUES (3, NULL), (NULL, 'x') INTERSECT VALUES (1, NULL), (2, NULL)
 intersect
  ├── columns: column1:1(int) column2:2(string)
  ├── left columns: column1:1(int) column2:2(string)
- ├── right columns: column1:3(int) column2:5(string)
+ ├── right columns: column1:3(int) column2:4(string)
  ├── values
  │    ├── columns: column1:1(int) column2:2(string)
  │    ├── tuple [type=tuple{int, string}]
@@ -885,27 +936,24 @@ intersect
  │         ├── cast: INT8 [type=int]
  │         │    └── null [type=unknown]
  │         └── const: 'x' [type=string]
- └── project
-      ├── columns: column2:5(string) column1:3(int!null)
-      ├── values
-      │    ├── columns: column1:3(int!null) column2:4(unknown)
-      │    ├── tuple [type=tuple{int, unknown}]
-      │    │    ├── const: 1 [type=int]
-      │    │    └── null [type=unknown]
-      │    └── tuple [type=tuple{int, unknown}]
-      │         ├── const: 2 [type=int]
+ └── values
+      ├── columns: column1:3(int!null) column2:4(string)
+      ├── tuple [type=tuple{int, string}]
+      │    ├── const: 1 [type=int]
+      │    └── cast: STRING [type=string]
       │         └── null [type=unknown]
-      └── projections
+      └── tuple [type=tuple{int, string}]
+           ├── const: 2 [type=int]
            └── cast: STRING [type=string]
-                └── variable: column2 [type=unknown]
+                └── null [type=unknown]
 
 build
 VALUES (NULL, NULL), (NULL, 'x') UNION ALL VALUES (1, NULL), (2, NULL)
 ----
 union-all
  ├── columns: column1:5(int) column2:6(string)
  ├── left columns: column1:7(int) column2:2(string)
- ├── right columns: column1:3(int) column2:8(string)
+ ├── right columns: column1:3(int) column2:4(string)
  ├── project
  │    ├── columns: column1:7(int) column2:2(string)
  │    ├── values
@@ -920,19 +968,16 @@ union-all
  │    └── projections
  │         └── cast: INT8 [type=int]
  │              └── variable: column1 [type=unknown]
- └── project
-      ├── columns: column2:8(string) column1:3(int!null)
-      ├── values
-      │    ├── columns: column1:3(int!null) column2:4(unknown)
-      │    ├── tuple [type=tuple{int, unknown}]
-      │    │    ├── const: 1 [type=int]
-      │    │    └── null [type=unknown]
-      │    └── tuple [type=tuple{int, unknown}]
-      │         ├── const: 2 [type=int]
+ └── values
+      ├── columns: column1:3(int!null) column2:4(string)
+      ├── tuple [type=tuple{int, string}]
+      │    ├── const: 1 [type=int]
+      │    └── cast: STRING [type=string]
       │         └── null [type=unknown]
-      └── projections
+      └── tuple [type=tuple{int, string}]
+           ├── const: 2 [type=int]
            └── cast: STRING [type=string]
-                └── variable: column2 [type=unknown]
+                └── null [type=unknown]
 
 build
 VALUES (NULL, NULL), (NULL, NULL) UNION ALL VALUES (NULL, NULL), (NULL, NULL)

pkg/sql/opt/optbuilder/union.go

@@ -29,6 +29,14 @@ func (b *Builder) buildUnion(
 	clause *tree.UnionClause, desiredTypes []*types.T, inScope *scope,
 ) (outScope *scope) {
 	leftScope := b.buildSelect(clause.Left, desiredTypes, inScope)
+	// Try to propagate types left-to-right, if we didn't already have desired
+	// types.
+	if len(desiredTypes) == 0 {
+		desiredTypes = make([]*types.T, len(leftScope.cols))
+		for i := range leftScope.cols {
+			desiredTypes[i] = leftScope.cols[i].typ
+		}
+	}
 	rightScope := b.buildSelect(clause.Right, desiredTypes, inScope)
 
 	// Remove any hidden columns, as they are not included in the Union.