Merge cockroachdb#39133

39133: exec: minor cleanup of planning of hash and merge joiners r=yuzefovich a=yuzefovich

A minor bug with setting up output column types with hash joiner is
fixed. Also, in some cases, we need to use a join hint to force
planning of hash joiner in the test.

Release note: None

Co-authored-by: Yahor Yuzefovich <yahor@cockroachlabs.com>

pkg/sql/distsqlrun/column_exec_setup.go

@@ -300,16 +300,16 @@ func newColOperator(
 			return result, err
 		}
 
-		columnTypes = make([]semtypes.T, len(leftTypes)+len(rightTypes))
-		copy(columnTypes, spec.Input[0].ColumnTypes)
-		copy(columnTypes[len(leftTypes):], spec.Input[1].ColumnTypes)
-
 		nLeftCols := uint32(len(leftTypes))
 		nRightCols := uint32(len(rightTypes))
 
 		leftOutCols := make([]uint32, 0)
 		rightOutCols := make([]uint32, 0)
 
+		// Note that we do not need a special treatment in case of LEFT SEMI and
+		// LEFT ANTI joins when setting up outCols because in such cases there will
+		// be a projection with post.OutputColumns already projecting out the right
+		// side.
 		if post.Projection {
 			for _, col := range post.OutputColumns {
 				if col < nLeftCols {
@@ -322,7 +322,6 @@ func newColOperator(
 			for i := uint32(0); i < nLeftCols; i++ {
 				leftOutCols = append(leftOutCols, i)
 			}
-
 			for i := uint32(0); i < nRightCols; i++ {
 				rightOutCols = append(rightOutCols, i)
 			}
@@ -345,6 +344,15 @@ func newColOperator(
 			return result, err
 		}
 
+		columnTypes = make([]semtypes.T, nLeftCols+nRightCols)
+		copy(columnTypes, spec.Input[0].ColumnTypes)
+		if core.HashJoiner.Type != sqlbase.JoinType_LEFT_SEMI {
+			// TODO(yuzefovich): update this conditional once LEFT ANTI is supported.
+			copy(columnTypes[nLeftCols:], spec.Input[1].ColumnTypes)
+		} else {
+			columnTypes = columnTypes[:nLeftCols]
+		}
+
 		if !core.HashJoiner.OnExpr.Empty() {
 			if core.HashJoiner.Type != sqlbase.JoinType_INNER {
 				return result, errors.Newf("can't plan non-inner hash join with on expressions")
@@ -377,25 +385,24 @@ func newColOperator(
 		leftOutCols := make([]uint32, 0, nLeftCols)
 		rightOutCols := make([]uint32, 0, nRightCols)
 
+		// Note that we do not need a special treatment in case of LEFT SEMI and
+		// LEFT ANTI joins when setting up outCols because in such cases there will
+		// be a projection with post.OutputColumns already projecting out the right
+		// side.
 		if post.Projection {
 			for _, col := range post.OutputColumns {
 				if col < nLeftCols {
 					leftOutCols = append(leftOutCols, col)
-				} else if core.MergeJoiner.Type != sqlbase.JoinType_LEFT_SEMI &&
-					core.MergeJoiner.Type != sqlbase.JoinType_LEFT_ANTI {
+				} else {
 					rightOutCols = append(rightOutCols, col-nLeftCols)
 				}
 			}
 		} else {
 			for i := uint32(0); i < nLeftCols; i++ {
 				leftOutCols = append(leftOutCols, i)
 			}
-
-			if core.MergeJoiner.Type != sqlbase.JoinType_LEFT_SEMI &&
-				core.MergeJoiner.Type != sqlbase.JoinType_LEFT_ANTI {
-				for i := uint32(0); i < nRightCols; i++ {
-					rightOutCols = append(rightOutCols, i)
-				}
+			for i := uint32(0); i < nRightCols; i++ {
+				rightOutCols = append(rightOutCols, i)
 			}
 		}
 

pkg/sql/distsqlrun/columnar_operators_test.go

@@ -382,6 +382,8 @@ func generateColumnOrdering(
 // comparison which can be either comparing a column from the left against a
 // column from the right or comparing a column from either side against a
 // constant.
+// TODO(yuzefovich): update this once LEFT SEMI or LEFT ANTI is supported with
+// ON expression.
 func generateOnExpr(rng *rand.Rand, nCols int, nEqCols int, maxNum int) distsqlpb.Expression {
 	var comparison string
 	r := rng.Float64()

pkg/sql/logictest/testdata/logic_test/exec_hash_join

@@ -1,7 +1,7 @@
 # LogicTest: local-vec
 
 # Test that the exec HashJoiner follows SQL NULL semantics for ON predicate
-# equivalence. The use of sorts here force the planning of merge join.
+# equivalence.
 
 statement ok
 CREATE TABLE  t1 (k INT PRIMARY KEY, v INT)
@@ -34,41 +34,41 @@ statement ok
 INSERT INTO c VALUES (1, 'a'), (1, 'b'), (2, 'c')
 
 query IIII
-SELECT * FROM t1 JOIN t2 ON t1.k = t2.x
+SELECT * FROM t1 INNER HASH JOIN t2 ON t1.k = t2.x ORDER BY 1
 ----
 0  4  0  5
 3  4  3  2
 
-query IIII rowsort
-SELECT * FROM a AS a1 JOIN a AS a2 ON a1.k = a2.v
+query IIII
+SELECT * FROM a AS a1 JOIN a AS a2 ON a1.k = a2.v ORDER BY 1
 ----
 0  1  2  0
 1  2  0  1
 2  0  1  2
 
-query IIII rowsort
-SELECT * FROM a AS a2 JOIN a AS a1 ON a1.k = a2.v
+query IIII
+SELECT * FROM a AS a2 JOIN a AS a1 ON a1.k = a2.v ORDER BY 1
 ----
 0  1  1  2
 1  2  2  0
 2  0  0  1
 
 query II
-SELECT t2.y, t1.v FROM t1 JOIN t2 ON t1.k = t2.x
+SELECT t2.y, t1.v FROM t1 INNER HASH JOIN t2 ON t1.k = t2.x ORDER BY 1 DESC
 ----
 5  4
 2  4
 
-query IIII rowsort
-SELECT * FROM t1 JOIN t2 ON t1.v = t2.x
+query IIII
+SELECT * FROM t1 JOIN t2 ON t1.v = t2.x ORDER BY 1
 ----
 0  4  4  6
 2  1  1  3
 3  4  4  6
 5  4  4  6
 
-query IIII rowsort
-SELECT * FROM t1 LEFT JOIN t2 ON t1.v = t2.x
+query IIII
+SELECT * FROM t1 LEFT JOIN t2 ON t1.v = t2.x ORDER BY 1
 ----
 -1  -1  NULL  NULL
 0   4   4     6
@@ -97,14 +97,14 @@ SELECT * FROM t1 FULL JOIN t2 ON t1.v = t2.x
 NULL  NULL  3     2
 NULL  NULL  0     5
 
-query IIT rowsort
-SELECT b.a, b.b, b.c FROM b JOIN a ON b.a = a.k AND a.v = b.b
+query IIT
+SELECT b.a, b.b, b.c FROM b JOIN a ON b.a = a.k AND a.v = b.b ORDER BY 3
 ----
 0  1  a
 0  1  d
 
 query ITI
-SELECT b.a, b.c, c.a FROM b JOIN c ON b.b = c.a AND b.c = c.b
+SELECT b.a, b.c, c.a FROM b JOIN c ON b.b = c.a AND b.c = c.b ORDER BY 2
 ----
 0  a  1
 2  b  1