planner: support hash v2 variable to control the join order

pkg/executor/BUILD.bazel

@@ -129,6 +129,7 @@ go_library(
         "//pkg/executor/internal/util",
         "//pkg/executor/internal/vecgroupchecker",
         "//pkg/executor/join",
+        "//pkg/executor/join/joinversion",
         "//pkg/executor/lockstats",
         "//pkg/executor/metrics",
         "//pkg/executor/sortexec",

pkg/executor/builder.go

@@ -48,6 +48,7 @@ import (
 	"github.com/pingcap/tidb/pkg/executor/internal/testutil"
 	"github.com/pingcap/tidb/pkg/executor/internal/vecgroupchecker"
 	"github.com/pingcap/tidb/pkg/executor/join"
+	"github.com/pingcap/tidb/pkg/executor/join/joinversion"
 	"github.com/pingcap/tidb/pkg/executor/lockstats"
 	executor_metrics "github.com/pingcap/tidb/pkg/executor/metrics"
 	"github.com/pingcap/tidb/pkg/executor/sortexec"
@@ -1748,7 +1749,7 @@ func (b *executorBuilder) buildHashJoinV2(v *plannercore.PhysicalHashJoin) exec.
 }
 
 func (b *executorBuilder) buildHashJoin(v *plannercore.PhysicalHashJoin) exec.Executor {
-	if b.ctx.GetSessionVars().UseHashJoinV2 && join.IsHashJoinV2Supported() && v.CanUseHashJoinV2() {
+	if b.ctx.GetSessionVars().UseHashJoinV2 && joinversion.IsHashJoinV2Supported() && v.CanUseHashJoinV2() {
 		return b.buildHashJoinV2(v)
 	}
 	leftExec := b.build(v.Children()[0])

pkg/executor/join/hash_join_v2.go

@@ -58,14 +58,6 @@ var (
 	HashJoinV2Strings = []string{DisableHashJoinV2, EnableHashJoinV2}
 )
 
-// IsHashJoinV2Supported return true if hash join v2 is supported in current env
-func IsHashJoinV2Supported() bool {
-	// sizeOfUintptr should always equal to sizeOfUnsafePointer, because according to golang's doc,
-	// a Pointer can be converted to an uintptr. Add this check here in case in the future go runtime
-	// change this
-	return !heapObjectsCanMove() && sizeOfUintptr >= sizeOfUnsafePointer
-}
-
 type hashTableContext struct {
 	// rowTables is used during split partition stage, each buildWorker has
 	// its own rowTable

pkg/executor/join/joinversion/join_version.go

@@ -16,8 +16,12 @@ package joinversion
 
 import (
 	"strings"
+	"unsafe"
 )
 
+//go:linkname heapObjectsCanMove runtime.heapObjectsCanMove
+func heapObjectsCanMove() bool
+
 const (
 	// HashJoinVersionLegacy means hash join v1
 	HashJoinVersionLegacy = "legacy"
@@ -43,3 +47,16 @@ func init() {
 func IsOptimizedVersion(hashJoinVersion string) bool {
 	return strings.ToLower(hashJoinVersion) == HashJoinVersionOptimized
 }
+
+const (
+	sizeOfUintptr       = int(unsafe.Sizeof(uintptr(0)))
+	sizeOfUnsafePointer = int(unsafe.Sizeof(unsafe.Pointer(nil)))
+)
+
+// IsHashJoinV2Supported return true if hash join v2 is supported in current env
+func IsHashJoinV2Supported() bool {
+	// sizeOfUintptr should always equal to sizeOfUnsafePointer, because according to golang's doc,
+	// a Pointer can be converted to an uintptr. Add this check here in case in the future go runtime
+	// change this
+	return !heapObjectsCanMove() && sizeOfUintptr >= sizeOfUnsafePointer
+}

pkg/planner/core/casetest/enforcempp/testdata/enforce_mpp_suite_out.json

@@ -1658,29 +1658,32 @@
       {
         "SQL": "EXPLAIN select /*+ HASH_JOIN_BUILD(s@sel_2) */  * from t where t.a not in (select s.a from s where t.c > s.c); -- 9. anti semi join, one join key + other condition + hint",
         "Plan": [
-          "TableReader_28 8000.00 root  MppVersion: 2, data:ExchangeSender_27",
-          "└─ExchangeSender_27 8000.00 mpp[tiflash]  ExchangeType: PassThrough",
-          "  └─HashJoin_22 8000.00 mpp[tiflash]  Null-aware anti semi join, left side:TableFullScan_23, equal:[eq(test.t.a, test.s.a)], other cond:gt(test.t.c, test.s.c)",
-          "    ├─ExchangeReceiver_26(Build) 10000.00 mpp[tiflash]  ",
-          "    │ └─ExchangeSender_25 10000.00 mpp[tiflash]  ExchangeType: Broadcast, Compression: FAST",
-          "    │   └─TableFullScan_24 10000.00 mpp[tiflash] table:s keep order:false, stats:pseudo",
-          "    └─TableFullScan_23(Probe) 10000.00 mpp[tiflash] table:t keep order:false, stats:pseudo"
+          "TableReader_30 8000.00 root  MppVersion: 2, data:ExchangeSender_29",
+          "└─ExchangeSender_29 8000.00 mpp[tiflash]  ExchangeType: PassThrough",
+          "  └─HashJoin_28 8000.00 mpp[tiflash]  Null-aware anti semi join, left side:TableFullScan_12, equal:[eq(test.t.a, test.s.a)], other cond:gt(test.t.c, test.s.c)",
+          "    ├─ExchangeReceiver_15(Build) 10000.00 mpp[tiflash]  ",
+          "    │ └─ExchangeSender_14 10000.00 mpp[tiflash]  ExchangeType: Broadcast, Compression: FAST",
+          "    │   └─TableFullScan_13 10000.00 mpp[tiflash] table:s keep order:false, stats:pseudo",
+          "    └─TableFullScan_12(Probe) 10000.00 mpp[tiflash] table:t keep order:false, stats:pseudo"
         ],
-        "Warn": null
+        "Warn": [
+          "[planner:1815]We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for anti semi join, please check the hint"
+        ]
       },
       {
         "SQL": "EXPLAIN select /*+ HASH_JOIN_BUILD(t@sel_1) */  * from t where t.a not in (select s.a from s where t.c > s.c); -- 10. anti semi join, one join key + other condition + hint",
         "Plan": [
-          "TableReader_28 8000.00 root  MppVersion: 2, data:ExchangeSender_27",
-          "└─ExchangeSender_27 8000.00 mpp[tiflash]  ExchangeType: PassThrough",
-          "  └─HashJoin_22 8000.00 mpp[tiflash]  Null-aware anti semi join, left side:TableFullScan_23, equal:[eq(test.t.a, test.s.a)], other cond:gt(test.t.c, test.s.c)",
-          "    ├─ExchangeReceiver_26(Build) 10000.00 mpp[tiflash]  ",
-          "    │ └─ExchangeSender_25 10000.00 mpp[tiflash]  ExchangeType: Broadcast, Compression: FAST",
-          "    │   └─TableFullScan_24 10000.00 mpp[tiflash] table:s keep order:false, stats:pseudo",
-          "    └─TableFullScan_23(Probe) 10000.00 mpp[tiflash] table:t keep order:false, stats:pseudo"
+          "TableReader_30 8000.00 root  MppVersion: 2, data:ExchangeSender_29",
+          "└─ExchangeSender_29 8000.00 mpp[tiflash]  ExchangeType: PassThrough",
+          "  └─HashJoin_28 8000.00 mpp[tiflash]  Null-aware anti semi join, left side:TableFullScan_12, equal:[eq(test.t.a, test.s.a)], other cond:gt(test.t.c, test.s.c)",
+          "    ├─ExchangeReceiver_15(Build) 10000.00 mpp[tiflash]  ",
+          "    │ └─ExchangeSender_14 10000.00 mpp[tiflash]  ExchangeType: Broadcast, Compression: FAST",
+          "    │   └─TableFullScan_13 10000.00 mpp[tiflash] table:s keep order:false, stats:pseudo",
+          "    └─TableFullScan_12(Probe) 10000.00 mpp[tiflash] table:t keep order:false, stats:pseudo"
         ],
         "Warn": [
           "[planner:1815]Some HASH_JOIN_BUILD and HASH_JOIN_PROBE hints cannot be utilized for MPP joins, please check the hints",
+          "[planner:1815]We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for anti semi join, please check the hint",
           "[planner:1815]Some HASH_JOIN_BUILD and HASH_JOIN_PROBE hints cannot be utilized for MPP joins, please check the hints"
         ]
       }

pkg/planner/core/casetest/join/join_test.go

@@ -35,9 +35,10 @@ func TestSemiJoinOrder(t *testing.T) {
 		"3 1",
 		"3 2",
 		"3 4")
+	tk.MustExec("set tidb_hash_join_version=optimized")
 	tk.MustQuery("select * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
 	tk.MustQuery("select /*+ HASH_JOIN_BUILD(t1) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
-	tk.MustQuery("select /*+ HASH_JOIN_BUILD(t2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
+	tk.MustQuery("select /*+ HASH_JOIN_BUILD(t2@sel_2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
 	tk.MustQuery("explain format = 'brief' select  /*+ HASH_JOIN_BUILD(t1) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(testkit.Rows(
 		"Sort 7992.00 root  test.t1.col0, test.t1.col1",
 		"└─HashJoin 7992.00 root  semi join, left side:TableReader, equal:[eq(test.t1.col0, test.t2.col0)]",
@@ -47,7 +48,8 @@ func TestSemiJoinOrder(t *testing.T) {
 		"  └─TableReader(Probe) 9990.00 root  data:Selection",
 		"    └─Selection 9990.00 cop[tikv]  not(isnull(test.t2.col0))",
 		"      └─TableFullScan 10000.00 cop[tikv] table:t2 keep order:false, stats:pseudo"))
-	tk.MustQuery("explain format = 'brief' select  /*+ HASH_JOIN_BUILD(t2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(testkit.Rows(
+	tk.MustQuery("show warnings").Check(testkit.Rows())
+	tk.MustQuery("explain format = 'brief' select  /*+ HASH_JOIN_BUILD(t2@sel_2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(testkit.Rows(
 		"Sort 7992.00 root  test.t1.col0, test.t1.col1",
 		"└─HashJoin 7992.00 root  semi join, left side:TableReader, equal:[eq(test.t1.col0, test.t2.col0)]",
 		"  ├─TableReader(Build) 9990.00 root  data:Selection",
@@ -56,4 +58,33 @@ func TestSemiJoinOrder(t *testing.T) {
 		"  └─TableReader(Probe) 9990.00 root  data:Selection",
 		"    └─Selection 9990.00 cop[tikv]  not(isnull(test.t1.col0))",
 		"      └─TableFullScan 10000.00 cop[tikv] table:t1 keep order:false, stats:pseudo"))
+	tk.MustQuery("show warnings").Check(testkit.Rows())
+	tk.MustExec("set tidb_hash_join_version=legacy")
+	tk.MustQuery("select * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
+	tk.MustQuery("select /*+ HASH_JOIN_BUILD(t1) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
+	tk.MustQuery("select /*+ HASH_JOIN_BUILD(t2@sel_2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(result)
+	tk.MustQuery("explain format = 'brief' select  /*+ HASH_JOIN_BUILD(t1) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(testkit.Rows(
+		"Sort 7992.00 root  test.t1.col0, test.t1.col1",
+		"└─HashJoin 7992.00 root  semi join, left side:TableReader, equal:[eq(test.t1.col0, test.t2.col0)]",
+		"  ├─TableReader(Build) 9990.00 root  data:Selection",
+		"  │ └─Selection 9990.00 cop[tikv]  not(isnull(test.t2.col0))",
+		"  │   └─TableFullScan 10000.00 cop[tikv] table:t2 keep order:false, stats:pseudo",
+		"  └─TableReader(Probe) 9990.00 root  data:Selection",
+		"    └─Selection 9990.00 cop[tikv]  not(isnull(test.t1.col0))",
+		"      └─TableFullScan 10000.00 cop[tikv] table:t1 keep order:false, stats:pseudo"))
+	tk.MustQuery("show warnings").Check(testkit.Rows(
+		"Warning 1815 We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for semi join, please check the hint",
+		"Warning 1815 We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for semi join, please check the hint"))
+	tk.MustQuery("explain format = 'brief' select  /*+ HASH_JOIN_BUILD(t2@sel_2) */ * from t1 where exists (select 1 from t2 where t1.col0 = t2.col0) order by t1.col0, t1.col1;").Check(testkit.Rows(
+		"Sort 7992.00 root  test.t1.col0, test.t1.col1",
+		"└─HashJoin 7992.00 root  semi join, left side:TableReader, equal:[eq(test.t1.col0, test.t2.col0)]",
+		"  ├─TableReader(Build) 9990.00 root  data:Selection",
+		"  │ └─Selection 9990.00 cop[tikv]  not(isnull(test.t2.col0))",
+		"  │   └─TableFullScan 10000.00 cop[tikv] table:t2 keep order:false, stats:pseudo",
+		"  └─TableReader(Probe) 9990.00 root  data:Selection",
+		"    └─Selection 9990.00 cop[tikv]  not(isnull(test.t1.col0))",
+		"      └─TableFullScan 10000.00 cop[tikv] table:t1 keep order:false, stats:pseudo"))
+	tk.MustQuery("show warnings").Check(testkit.Rows(
+		"Warning 1815 We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for semi join, please check the hint",
+		"Warning 1815 We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for semi join, please check the hint"))
 }

pkg/planner/core/exhaust_physical_plans.go

@@ -22,6 +22,7 @@ import (
 
 	"github.com/pingcap/errors"
 	"github.com/pingcap/failpoint"
+	"github.com/pingcap/tidb/pkg/executor/join/joinversion"
 	"github.com/pingcap/tidb/pkg/expression"
 	"github.com/pingcap/tidb/pkg/expression/aggregation"
 	"github.com/pingcap/tidb/pkg/kv"
@@ -371,6 +372,56 @@ func shouldSkipHashJoin(p *logicalop.LogicalJoin) bool {
 	return (p.PreferJoinType&h.PreferNoHashJoin) > 0 || (p.SCtx().GetSessionVars().DisableHashJoin)
 }
 
+func isGAForHashJoinV2(joinType logicalop.JoinType, leftJoinKeys []*expression.Column, isNullEQ []bool, leftNAJoinKeys []*expression.Column) bool {
+	// nullaware join
+	if len(leftNAJoinKeys) > 0 {
+		return false
+	}
+	// cross join
+	if len(leftJoinKeys) == 0 {
+		return false
+	}
+	// join with null equal condition
+	for _, value := range isNullEQ {
+		if value {
+			return false
+		}
+	}
+	switch joinType {
+	case logicalop.LeftOuterJoin, logicalop.RightOuterJoin, logicalop.InnerJoin:
+		return true
+	default:
+		return false
+	}
+}
+
+// CanUseHashJoinV2 returns true if current join is supported by hash join v2
+func canUseHashJoinV2(joinType logicalop.JoinType, leftJoinKeys []*expression.Column, isNullEQ []bool, leftNAJoinKeys []*expression.Column) bool {
+	if !isGAForHashJoinV2(joinType, leftJoinKeys, isNullEQ, leftNAJoinKeys) && !joinversion.UseHashJoinV2ForNonGAJoin {
+		return false
+	}
+	switch joinType {
+	case logicalop.LeftOuterJoin, logicalop.RightOuterJoin, logicalop.InnerJoin, logicalop.LeftOuterSemiJoin, logicalop.SemiJoin, logicalop.AntiSemiJoin:
+		// null aware join is not supported yet
+		if len(leftNAJoinKeys) > 0 {
+			return false
+		}
+		// cross join is not supported
+		if len(leftJoinKeys) == 0 {
+			return false
+		}
+		// NullEQ is not supported yet
+		for _, value := range isNullEQ {
+			if value {
+				return false
+			}
+		}
+		return true
+	default:
+		return false
+	}
+}
+
 func getHashJoins(p *logicalop.LogicalJoin, prop *property.PhysicalProperty) (joins []base.PhysicalPlan, forced bool) {
 	if !prop.IsSortItemEmpty() { // hash join doesn't promise any orders
 		return
@@ -383,17 +434,28 @@ func getHashJoins(p *logicalop.LogicalJoin, prop *property.PhysicalProperty) (jo
 		forceLeftToBuild = false
 		forceRightToBuild = false
 	}
-
 	joins = make([]base.PhysicalPlan, 0, 2)
 	switch p.JoinType {
 	case logicalop.SemiJoin, logicalop.AntiSemiJoin:
-		if !forceLeftToBuild {
-			joins = append(joins, getHashJoin(p, prop, 1, false))
-		} else if !forceRightToBuild {
-			joins = append(joins, getHashJoin(p, prop, 1, true))
+		leftJoinKeys, _, isNullEQ, _ := p.GetJoinKeys()
+		leftNAJoinKeys, _ := p.GetNAJoinKeys()
+		if p.SCtx().GetSessionVars().UseHashJoinV2 && joinversion.IsHashJoinV2Supported() && canUseHashJoinV2(p.JoinType, leftJoinKeys, isNullEQ, leftNAJoinKeys) {
+			if !forceLeftToBuild {
+				joins = append(joins, getHashJoin(p, prop, 1, false))
+			} else if !forceRightToBuild {
+				joins = append(joins, getHashJoin(p, prop, 1, true))
+			} else {
+				joins = append(joins, getHashJoin(p, prop, 1, false))
+				joins = append(joins, getHashJoin(p, prop, 1, true))
+			}
 		} else {
 			joins = append(joins, getHashJoin(p, prop, 1, false))
-			joins = append(joins, getHashJoin(p, prop, 1, true))
+			if forceLeftToBuild || forceRightToBuild {
+				// Do not support specifying the build and probe side for semi join.
+				p.SCtx().GetSessionVars().StmtCtx.SetHintWarning(fmt.Sprintf("We can't use the HASH_JOIN_BUILD or HASH_JOIN_PROBE hint for %s, please check the hint", p.JoinType))
+				forceLeftToBuild = false
+				forceRightToBuild = false
+			}
 		}
 	case logicalop.LeftOuterSemiJoin, logicalop.AntiLeftOuterSemiJoin:
 		joins = append(joins, getHashJoin(p, prop, 1, false))

pkg/planner/core/physical_plans.go

@@ -21,7 +21,6 @@ import (
 	"unsafe"
 
 	"github.com/pingcap/errors"
-	"github.com/pingcap/tidb/pkg/executor/join/joinversion"
 	"github.com/pingcap/tidb/pkg/expression"
 	"github.com/pingcap/tidb/pkg/expression/aggregation"
 	"github.com/pingcap/tidb/pkg/kv"
@@ -1490,54 +1489,9 @@ type PhysicalHashJoin struct {
 	runtimeFilterList []*RuntimeFilter `plan-cache-clone:"must-nil"` // plan with runtime filter is not cached
 }
 
-func (p *PhysicalHashJoin) isGAForHashJoinV2() bool {
-	// nullaware join
-	if len(p.LeftNAJoinKeys) > 0 {
-		return false
-	}
-	// cross join
-	if len(p.LeftJoinKeys) == 0 {
-		return false
-	}
-	// join with null equal condition
-	for _, value := range p.IsNullEQ {
-		if value {
-			return false
-		}
-	}
-	switch p.JoinType {
-	case logicalop.LeftOuterJoin, logicalop.RightOuterJoin, logicalop.InnerJoin:
-		return true
-	default:
-		return false
-	}
-}
-
 // CanUseHashJoinV2 returns true if current join is supported by hash join v2
 func (p *PhysicalHashJoin) CanUseHashJoinV2() bool {
-	if !p.isGAForHashJoinV2() && !joinversion.UseHashJoinV2ForNonGAJoin {
-		return false
-	}
-	switch p.JoinType {
-	case logicalop.LeftOuterJoin, logicalop.RightOuterJoin, logicalop.InnerJoin, logicalop.LeftOuterSemiJoin, logicalop.SemiJoin, logicalop.AntiSemiJoin, logicalop.AntiLeftOuterSemiJoin:
-		// null aware join is not supported yet
-		if len(p.LeftNAJoinKeys) > 0 {
-			return false
-		}
-		// cross join is not supported
-		if len(p.LeftJoinKeys) == 0 {
-			return false
-		}
-		// NullEQ is not supported yet
-		for _, value := range p.IsNullEQ {
-			if value {
-				return false
-			}
-		}
-		return true
-	default:
-		return false
-	}
+	return canUseHashJoinV2(p.JoinType, p.LeftJoinKeys, p.IsNullEQ, p.LeftNAJoinKeys)
 }
 
 // Clone implements op.PhysicalPlan interface.