[Plan] Deduplicate identical UNION parts (#4552)

pkg/sql/plan/query_builder.go

@@ -661,6 +661,14 @@ func (builder *QueryBuilder) buildUnion(stmt *tree.Select, ctx *BindContext, isR
 		return 0, err
 	}
 
+	if len(selectStmts) == 1 {
+		if slt, ok := selectStmts[0].(*tree.Select); ok {
+			return builder.buildSelect(slt, ctx, false)
+		} else {
+			return builder.buildSelect(&tree.Select{Select: selectStmts[0]}, ctx, false)
+		}
+	}
+
 	// build selects
 	var projectTypList [][]types.Type
 	selectStmtLength := len(selectStmts)

pkg/sql/plan/utils.go

@@ -515,8 +515,22 @@ func getUnionSelects(stmt *tree.UnionClause, selects *[]tree.Statement, unionTyp
 	// right is not UNION allways
 	switch rightStmt := stmt.Right.(type) {
 	case *tree.SelectClause:
+		if stmt.Type == tree.UNION && !stmt.All {
+			rightStr := tree.String(rightStmt, dialect.MYSQL)
+			if len(*selects) == 1 && tree.String((*selects)[0], dialect.MYSQL) == rightStr {
+				return nil
+			}
+		}
+
 		*selects = append(*selects, rightStmt)
 	case *tree.ParenSelect:
+		if stmt.Type == tree.UNION && !stmt.All {
+			rightStr := tree.String(rightStmt.Select, dialect.MYSQL)
+			if len(*selects) == 1 && tree.String((*selects)[0], dialect.MYSQL) == rightStr {
+				return nil
+			}
+		}
+
 		*selects = append(*selects, rightStmt.Select)
 	default:
 		return errors.New(errno.SQLStatementNotYetComplete, fmt.Sprintf("unexpected statement in union2: '%v'", tree.String(rightStmt, dialect.MYSQL)))
@@ -531,13 +545,13 @@ func getUnionSelects(stmt *tree.UnionClause, selects *[]tree.Statement, unionTyp
 		}
 	case tree.INTERSECT:
 		if stmt.All {
-			return errors.New("", "INTERSECT ALL clause will support in future version.")
+			return errors.New("", "INTERSECT ALL clause will be supported in future version.")
 		} else {
 			*unionTypes = append(*unionTypes, plan.Node_INTERSECT)
 		}
 	case tree.EXCEPT, tree.UT_MINUS:
 		if stmt.All {
-			return errors.New("", "EXCEPT/MINUS ALL clause will support in future version.")
+			return errors.New("", "EXCEPT/MINUS ALL clause will be supported in future version.")
 		} else {
 			*unionTypes = append(*unionTypes, plan.Node_MINUS)
 		}

test/cases/subquery/subquery-with-any.sql

@@ -78,12 +78,8 @@ SELECT a FROM t1 WHERE (1,2) <> ALL (SELECT a FROM t1 WHERE b = 2);
 
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 WHERE b = 2);
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 HAVING a = 2);
--- @bvt:issue#4354
--- @bvt:issue#3419
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 WHERE b = 2 UNION SELECT a,1 FROM t1 WHERE b = 2);
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 HAVING a = 2 UNION SELECT a,1 FROM t1 HAVING a = 2);
--- @bvt:issue
--- @bvt:issue
 
 DROP TABLE IF EXISTS t1;
 CREATE TABLE t1 ( a double, b double );
@@ -323,13 +319,9 @@ DROP TABLE IF EXISTS t1;
 CREATE TABLE t1( a INT );
 INSERT INTO t1 VALUES (1),(2);
 CREATE TABLE t2( a INT, b INT );
--- @bvt:issue#4354
--- @bvt:issue#3419
 SELECT * FROM t1 WHERE a = ANY ( SELECT 1 UNION ( SELECT 1 UNION SELECT 1 ) );
 SELECT * FROM t1 WHERE a = ANY ( ( SELECT 1 UNION SELECT 1 )  UNION SELECT 1 );
 SELECT * FROM t1 WHERE a = ANY ( SELECT 1 UNION SELECT 1 UNION SELECT 1 );
--- @bvt:issue
--- @bvt:issue
 
 DROP TABLE IF EXISTS t1;
 DROP TABLE IF EXISTS t2;

test/cases/subquery/subquery-with-exists.sql

@@ -286,16 +286,12 @@ drop table if exists t1;
 -- @label:bvt
 drop table if exists t1;
 CREATE TABLE t1 (i INT);
--- @bvt:issue#4354
--- @bvt:issue#3419
 SELECT * FROM t1 WHERE NOT EXISTS
   (
    (SELECT i FROM t1) UNION
    (SELECT i FROM t1)
   );
 SELECT * FROM t1 WHERE NOT EXISTS (((SELECT i FROM t1) UNION (SELECT i FROM t1)));
--- @bvt:issue
--- @bvt:issue
 
 drop table if exists t1;
 drop table if exists t2;
@@ -341,11 +337,7 @@ CREATE TABLE t2 (a INT);
 INSERT INTO t1 VALUES (1),(2);
 INSERT INTO t2 VALUES (1),(2);
 SELECT 2 FROM t1 WHERE EXISTS ((SELECT 1 FROM t2 WHERE t1.a=t2.a));
--- @bvt:issue#4354
--- @bvt:issue#3419
 SELECT 2 FROM t1 WHERE EXISTS ((SELECT 1 FROM t2 WHERE t1.a=t2.a) UNION (SELECT 1 FROM t2 WHERE t1.a = t2.a));
--- @bvt:issue
--- @bvt:issue
 
 drop table if exists t1;
 drop table if exists t2;

test/cases/subquery/subquery-with-in.sql

@@ -199,15 +199,11 @@ insert into t3 values (NULL), (NULL);
 select a, oref, a in (select max(ie) from t1 where oref=t2.oref group by grp) Z from t2;
 -- @bvt:issue
 select a, oref from t2 where a in (select max(ie) from t1 where oref=t2.oref group by grp);
--- @bvt:issue#4354
--- @bvt:issue#3419
+-- @bvt:issue#3312
 select a, oref, a in (
   select max(ie) from t1 where oref=t2.oref group by grp union
   select max(ie) from t1 where oref=t2.oref group by grp
   ) Z from t2;
--- @bvt:issue
--- @bvt:issue
--- @bvt:issue#3312
 select a in (select max(ie) from t1 where oref=4 group by grp) from t3;
 -- @bvt:issue
 
@@ -619,13 +615,9 @@ INSERT INTO t1 VALUES (1),(2);
 CREATE TABLE t2( a INT, b INT );
 SELECT * FROM t2 WHERE (a, b) IN (SELECT a, b FROM t2);
 -- error
--- @bvt:issue#4354
--- @bvt:issue#3419
 SELECT * FROM t1 WHERE a IN    ( SELECT 1 UNION ( SELECT 1 UNION SELECT 1 ) );
 SELECT * FROM t1 WHERE a IN    ( ( SELECT 1 UNION SELECT 1 )  UNION SELECT 1 );
 SELECT * FROM t1 WHERE a IN    ( SELECT 1 UNION SELECT 1 UNION SELECT 1 );
--- @bvt:issue
--- @bvt:issue
 
 DROP TABLE IF EXISTS t1;
 drop table if exists t2;
@@ -876,12 +868,8 @@ create table t1 (a int, b int, primary key (a));
 insert into t1 values (1,1), (3,1),(100,1);
 create table t2 (a int, b int);
 insert into t2 values (1,1),(2,1),(NULL,1),(NULL,0);
--- @bvt:issue#4354
--- @bvt:issue#3419
-select a,b, a in (select a from t1 where t1.b = t2.b union select a from t1 where t1.b = t2.b) Z from t2 ;
--- @bvt:issue
--- @bvt:issue
 -- @bvt:issue#3312
+select a,b, a in (select a from t1 where t1.b = t2.b union select a from t1 where t1.b = t2.b) Z from t2 ;
 select a,b, a in (select a from t1 where t1.b = t2.b) Z from t2 ;
 -- @bvt:issue
 

test/result/subquery/subquery-with-any.result

@@ -138,9 +138,10 @@ a
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 HAVING a = 2);
 column "t1.a" must appear in the GROUP BY clause or be used in an aggregate function
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 WHERE b = 2 UNION SELECT a,1 FROM t1 WHERE b = 2);
-'union' will be supported in future version.
+a
+2
 SELECT a FROM t1 WHERE (a,1) = ANY (SELECT a,1 FROM t1 HAVING a = 2 UNION SELECT a,1 FROM t1 HAVING a = 2);
-'union' will be supported in future version.
+column "t1.a" must appear in the GROUP BY clause or be used in an aggregate function
 DROP TABLE IF EXISTS t1;
 CREATE TABLE t1 ( a double, b double );
 INSERT INTO t1 VALUES (1,1),(2,2),(3,3);
@@ -474,11 +475,14 @@ CREATE TABLE t1( a INT );
 INSERT INTO t1 VALUES (1),(2);
 CREATE TABLE t2( a INT, b INT );
 SELECT * FROM t1 WHERE a = ANY ( SELECT 1 UNION ( SELECT 1 UNION SELECT 1 ) );
-'union' will be supported in future version.
+a
+1
 SELECT * FROM t1 WHERE a = ANY ( ( SELECT 1 UNION SELECT 1 )  UNION SELECT 1 );
-'union' will be supported in future version.
+a
+1
 SELECT * FROM t1 WHERE a = ANY ( SELECT 1 UNION SELECT 1 UNION SELECT 1 );
-'union' will be supported in future version.
+a
+1
 DROP TABLE IF EXISTS t1;
 DROP TABLE IF EXISTS t2;
 CREATE TABLE t1(i INT);

test/result/subquery/subquery-with-exists.result

@@ -395,9 +395,9 @@ SELECT * FROM t1 WHERE NOT EXISTS
 (SELECT i FROM t1) UNION
 (SELECT i FROM t1)
 );
-'union' will be supported in future version.
+i
 SELECT * FROM t1 WHERE NOT EXISTS (((SELECT i FROM t1) UNION (SELECT i FROM t1)));
-'union' will be supported in future version.
+i
 drop table if exists t1;
 drop table if exists t2;
 drop table if exists t3;
@@ -436,7 +436,9 @@ SELECT 2 FROM t1 WHERE EXISTS ((SELECT 1 FROM t2 WHERE t1.a=t2.a));
 2
 2
 SELECT 2 FROM t1 WHERE EXISTS ((SELECT 1 FROM t2 WHERE t1.a=t2.a) UNION (SELECT 1 FROM t2 WHERE t1.a = t2.a));
-'union' will be supported in future version.
+2
+2
+2
 drop table if exists t1;
 drop table if exists t2;
 drop table if exists t1;

test/result/subquery/subquery-with-in.result

@@ -740,11 +740,14 @@ CREATE TABLE t2( a INT, b INT );
 SELECT * FROM t2 WHERE (a, b) IN (SELECT a, b FROM t2);
 a	b
 SELECT * FROM t1 WHERE a IN    ( SELECT 1 UNION ( SELECT 1 UNION SELECT 1 ) );
-'union' will be supported in future version.
+a
+1
 SELECT * FROM t1 WHERE a IN    ( ( SELECT 1 UNION SELECT 1 )  UNION SELECT 1 );
-'union' will be supported in future version.
+a
+1
 SELECT * FROM t1 WHERE a IN    ( SELECT 1 UNION SELECT 1 UNION SELECT 1 );
-'union' will be supported in future version.
+a
+1
 DROP TABLE IF EXISTS t1;
 drop table if exists t2;
 drop table if exists t3;