executor: fix data race in hash join v2 spill (#57519)

ref #57450

pkg/executor/join/hash_join_v2.go

@@ -504,11 +504,6 @@ func (b *BuildWorkerV2) splitPartitionAndAppendToRowTableForRestore(inDisk *chun
 		}
 	}()
 
-	partitionNumber := b.HashJoinCtx.partitionNumber
-	hashJoinCtx := b.HashJoinCtx
-
-	b.builder = createRowTableBuilder(b.BuildKeyColIdx, hashJoinCtx.BuildKeyTypes, partitionNumber, b.HasNullableKey, hashJoinCtx.BuildFilter != nil, hashJoinCtx.needScanRowTableAfterProbeDone)
-
 	hasErr := false
 	chunkNum := inDisk.NumChunks()
 	for i := 0; i < chunkNum; i++ {
@@ -544,10 +539,6 @@ func (b *BuildWorkerV2) splitPartitionAndAppendToRowTable(typeCtx types.Context,
 			setMaxValue(&b.HashJoinCtx.stats.maxPartitionData, cost)
 		}
 	}()
-	partitionNumber := b.HashJoinCtx.partitionNumber
-	hashJoinCtx := b.HashJoinCtx
-
-	b.builder = createRowTableBuilder(b.BuildKeyColIdx, hashJoinCtx.BuildKeyTypes, partitionNumber, b.HasNullableKey, hashJoinCtx.BuildFilter != nil, hashJoinCtx.needScanRowTableAfterProbeDone)
 
 	hasErr := false
 	for chk := range srcChkCh {
@@ -1263,6 +1254,11 @@ func (e *HashJoinV2Exec) fetchAndBuildHashTableImpl(ctx context.Context) {
 
 	// doneCh is used by the consumer(splitAndAppendToRowTable) to info the producer(fetchBuildSideRows) that the consumer meet error and stop consume data
 	doneCh := make(chan struct{}, e.Concurrency)
+	// init builder, todo maybe the builder can be reused during the whole life cycle of the executor
+	hashJoinCtx := e.HashJoinCtxV2
+	for _, worker := range e.BuildWorkers {
+		worker.builder = createRowTableBuilder(worker.BuildKeyColIdx, hashJoinCtx.BuildKeyTypes, hashJoinCtx.partitionNumber, worker.HasNullableKey, hashJoinCtx.BuildFilter != nil, hashJoinCtx.needScanRowTableAfterProbeDone)
+	}
 	srcChkCh, waitForController := e.fetchBuildSideRows(ctx, fetcherAndWorkerSyncer, wg, errCh, doneCh)
 	e.splitAndAppendToRowTable(srcChkCh, waitForController, fetcherAndWorkerSyncer, wg, errCh, doneCh)
 	success := waitJobDone(wg, errCh)