godfreyhe commented on a change in pull request #18707:
URL: https://github.com/apache/flink/pull/18707#discussion_r804709827



##########
File path: 
flink-table/flink-table-planner/src/main/java/org/apache/flink/table/planner/plan/nodes/exec/processor/ForwardHashExchangeProcessor.java
##########
@@ -0,0 +1,171 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.flink.table.planner.plan.nodes.exec.processor;
+
+import org.apache.flink.table.api.TableException;
+import org.apache.flink.table.planner.plan.nodes.exec.ExecEdge;
+import org.apache.flink.table.planner.plan.nodes.exec.ExecNode;
+import org.apache.flink.table.planner.plan.nodes.exec.ExecNodeGraph;
+import org.apache.flink.table.planner.plan.nodes.exec.InputProperty;
+import org.apache.flink.table.planner.plan.nodes.exec.batch.BatchExecExchange;
+import 
org.apache.flink.table.planner.plan.nodes.exec.batch.BatchExecMultipleInput;
+import 
org.apache.flink.table.planner.plan.nodes.exec.common.CommonExecExchange;
+import org.apache.flink.table.planner.plan.nodes.exec.stream.StreamExecNode;
+import 
org.apache.flink.table.planner.plan.nodes.exec.visitor.AbstractExecNodeExactlyOnceVisitor;
+import org.apache.flink.table.planner.plan.nodes.exec.visitor.ExecNodeVisitor;
+import org.apache.flink.table.types.logical.RowType;
+
+import java.util.ArrayList;
+import java.util.Collections;
+import java.util.List;
+
+import static org.apache.flink.util.Preconditions.checkArgument;
+
+/**
+ * A {@link ExecNodeGraphProcessor} which finds all {@link ExecNode}s that 
require hash
+ * distribution, but its input has no hash Exchange node. Currently, the hash 
distribution is
+ * satisfied by its none exchange inputs with FORWARD partitioner (the node 
and its input has the
+ * same parallelism). Once the parallelism is changed, the FORWARD behavior 
will be broken, and the
+ * result will be wrong.
+ *
+ * <p>In order to meet the needs of flexible parallelism changing, a special 
{@link
+ * BatchExecExchange} (with KEEP_INPUT_AS_IS distribution flag) will be added 
for the {@link
+ * ExecNode} as its input. And then the runtime will decide which partitioner 
can be used when
+ * adaptive scheduler is enabled: FORWARD partitioner if the parallelism is 
the same, or HASH

Review comment:
       updated




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Reply via email to