fix(query): reuse the left's column binding in union as possible (#18053)

* chore(query): reuse the left's column binding in union asas possible

* chore(query): reuse the left's column binding in union asas possible

* chore(query): reuse the left's column binding in union asas possible

* Update tests/sqllogictests/suites/query/union.test

Co-authored-by: Copilot <[email protected]>

* chore(query): reuse the left's column binding in union asas possible

* chore(query): reuse the left's column binding in union asas possible

* chore(query): reuse the left's column binding in union asas possible

---------

Co-authored-by: Bohu <[email protected]>
Co-authored-by: Copilot <[email protected]>

Author: 3 people

src/query/sql/src/planner/optimizer/optimizers/rule/union_rules/rule_eliminate_union.rs

@@ -72,6 +72,17 @@ impl Rule for RuleEliminateUnion {
 
     fn apply(&self, s_expr: &SExpr, state: &mut TransformResult) -> Result<()> {
         let union: UnionAll = s_expr.plan().clone().try_into()?;
+
+        // Need to check that union's output indexes are the same as left child's output indexes
+        // currently this is always !false now, so the following codes are not necessary
+        if !union
+            .left_outputs
+            .iter()
+            .all(|(idx, _)| union.output_indexes.contains(idx))
+        {
+            return Ok(());
+        }
+
         let left_child = s_expr.child(0)?;
         let right_child = s_expr.child(1)?;
 

tests/sqllogictests/suites/mode/standalone/explain/union.test

@@ -269,7 +269,11 @@ settings (ddl_column_type_nullable=0) create table t2 as select number as b from
 query T
 explain select * from t1 where t1.a < 0 union all select a from t2 join t1 on t1.a = t2.b where a <0;
 ----
-EmptyResultScan
+UnionAll
+├── output columns: [a (#3)]
+├── estimated rows: 0.00
+├── EmptyResultScan
+└── EmptyResultScan
 
 
 query T
@@ -295,21 +299,29 @@ UnionAll
 query T
 explain select * from t1 union all select * from t2 where t2.b < 0;
 ----
-TableScan
-├── table: default.default.t1
-├── output columns: []
-├── read rows: 10000
-├── read size: 0
-├── partitions total: 1
-├── partitions scanned: 1
-├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
-├── push downs: [filters: [], limit: NONE]
-└── estimated rows: 10000.00
+UnionAll
+├── output columns: [a (#2)]
+├── estimated rows: 10000.00
+├── TableScan
+│   ├── table: default.default.t1
+│   ├── output columns: [a (#0)]
+│   ├── read rows: 10000
+│   ├── read size: 10.59 KiB
+│   ├── partitions total: 1
+│   ├── partitions scanned: 1
+│   ├── pruning stats: [segments: <range pruning: 1 to 1>, blocks: <range pruning: 1 to 1>]
+│   ├── push downs: [filters: [], limit: NONE]
+│   └── estimated rows: 10000.00
+└── EmptyResultScan
 
 query T
 explain select * from t1 where t1.a < 0 union all select * from t2 where t2.b < 0;
 ----
-EmptyResultScan
+UnionAll
+├── output columns: [a (#2)]
+├── estimated rows: 0.00
+├── EmptyResultScan
+└── EmptyResultScan
 
 statement ok
 drop table t1;

tests/sqllogictests/suites/query/union.test

@@ -346,7 +346,19 @@ statement ok
 unset max_set_operator_count;
 
 query I
-select * from numbers(100) union all select * from numbers(100) ignore_result;
+select sum(number) as number, number* 3 as t   from numbers(100) group by t union all select  number, number* 3 as t from numbers(1000) where number < 0 ignore_result;
+----
+
+query I
+select t, sum(number) as number from (select sum(number) as number, number* 3 as t   from numbers(100) group by t union all select  number, number* 3 as t from numbers(1000) where number < 0) group by t ignore_result;
+----
+
+query I
+select  number::int32 from numbers(100) union all select number::int from numbers(100) where number < 0 ignore_result;
+----
+
+query I
+select number::int32 from numbers(100) union all select number::int from numbers(100) where number > 0 ignore_result;
 ----
 
 statement ok