[
https://issues.apache.org/jira/browse/FLINK-3179?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15183019#comment-15183019
]
ASF GitHub Bot commented on FLINK-3179:
---------------------------------------
Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/1553#discussion_r55205324
--- Diff:
flink-optimizer/src/main/java/org/apache/flink/optimizer/operators/GroupReduceWithCombineProperties.java
---
@@ -126,12 +117,66 @@ public SingleInputPlanNode instantiate(Channel in,
SingleInputNode node) {
}
toReducer.setLocalStrategy(LocalStrategy.COMBININGSORT,
in.getLocalStrategyKeys(),
in.getLocalStrategySortOrder());
-
return new SingleInputPlanNode(node, "Reduce
("+node.getOperator().getName()+")",
toReducer, DriverStrategy.SORTED_GROUP_REDUCE, this.keyList);
}
}
+ private SingleInputPlanNode injectCombinerBeforPartitioner(Channel in,
SingleInputNode node) {
+ // Inject a combiner before the partition node
+ Channel channelWithPartitionSrc = new Channel(in.getSource());
+ GroupReduceNode combinerNode = ((GroupReduceNode)
node).getCombinerUtilityNode();
+ combinerNode.setParallelism(in.getSource().getParallelism());
+ if(in.getSource().getInputs().iterator().hasNext()) {
+ Channel oldChannelToPartitioner =
channelWithPartitionSrc.getSource().getInputs().iterator().next();
+ Channel toCombiner = new
Channel(oldChannelToPartitioner.getSource());
+ // A combiner plan node is created with the channel as input
that has the partitioner as the target
+ toCombiner.setShipStrategy(ShipStrategyType.FORWARD,
DataExchangeMode.PIPELINED);
+ SingleInputPlanNode combiner = new
SingleInputPlanNode(combinerNode, "Combine("+node.getOperator()
+ .getName()+")", toCombiner,
DriverStrategy.SORTED_GROUP_COMBINE);
+ setCombinerProperties(in, oldChannelToPartitioner, combiner);
+
+ Channel toPartitioner = new Channel(combiner);
+ // Set the actual partitioner node's strategy key and
strategy order
+
toPartitioner.setShipStrategy(oldChannelToPartitioner.getShipStrategy(),
oldChannelToPartitioner.getShipStrategyKeys(),
+
oldChannelToPartitioner.getShipStrategySortOrder(),
oldChannelToPartitioner.getDataExchangeMode());
+ // Create the partition single input plan node from the
existing partition node
+ PlanNode partitionplanNode = in.getSource().getPlanNode();
+ SingleInputPlanNode partition = new
SingleInputPlanNode(in.getSource().getOptimizerNode(),
partitionplanNode.getNodeName(),
+ toPartitioner, partitionplanNode.getDriverStrategy());
+ partition.setCosts(partitionplanNode.getNodeCosts());
+
partition.initProperties(partitionplanNode.getGlobalProperties(),
partitionplanNode.getLocalProperties());
+ // Create a reducer such that the input of the reducer is the
partition node
+ Channel toReducer = new Channel(partition);
+ toReducer.setShipStrategy(in.getShipStrategy(),
in.getShipStrategyKeys(),
+ in.getShipStrategySortOrder(), in.getDataExchangeMode());
+ return getReducerSingleInputPlanNode(toReducer, node);
+ } else {
+ return getReducerSingleInputPlanNode(in, node);
--- End diff --
in.getSource is the partitioner an should thus have exactly one input. If
this is not the case, something is wrong and we should throw a
`CompilerException`.
> Combiner is not injected if Reduce or GroupReduce input is explicitly
> partitioned
> ---------------------------------------------------------------------------------
>
> Key: FLINK-3179
> URL: https://issues.apache.org/jira/browse/FLINK-3179
> Project: Flink
> Issue Type: Bug
> Components: Optimizer
> Affects Versions: 0.10.1
> Reporter: Fabian Hueske
> Assignee: ramkrishna.s.vasudevan
> Priority: Critical
> Fix For: 1.0.0, 0.10.2
>
>
> The optimizer does not inject a combiner if the input of a Reducer or
> GroupReducer is explicitly partitioned as in the following example
> {code}
> DataSet<Tuple2<String,Integer>> words = ...
> DataSet<Tuple2<String,Integer>> counts = words
> .partitionByHash(0)
> .groupBy(0)
> .sum(1);
> {code}
> Explicit partitioning can be useful to enforce partitioning on a subset of
> keys or to use a different partitioning method (custom or range partitioning).
> This issue should be fixed by changing the {{instantiate()}} methods of the
> {{ReduceProperties}} and {{GroupReduceWithCombineProperties}} classes such
> that a combine is injected in front of a {{PartitionPlanNode}} if it is the
> input of a Reduce or GroupReduce operator. This should only happen, if the
> Reducer is the only successor of the Partition operator.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
