flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From greghogan <...@git.apache.org>
Subject [GitHub] flink pull request #1517: [FLINK-3477] [runtime] Add hash-based combine stra...
Date Mon, 27 Jun 2016 19:36:43 GMT
Github user greghogan commented on a diff in the pull request:

    https://github.com/apache/flink/pull/1517#discussion_r68641370
  
    --- Diff: flink-optimizer/src/main/java/org/apache/flink/optimizer/dag/ReduceNode.java
---
    @@ -45,10 +46,28 @@ public ReduceNode(ReduceOperatorBase<?, ?> operator) {
     			// case of a key-less reducer. force a parallelism of 1
     			setParallelism(1);
     		}
    -		
    -		OperatorDescriptorSingle props = this.keys == null ?
    -			new AllReduceProperties() :
    -			new ReduceProperties(this.keys, operator.getCustomPartitioner());
    +
    +		OperatorDescriptorSingle props;
    +
    +		if (this.keys == null) {
    +			props = new AllReduceProperties();
    +		} else {
    +			DriverStrategy combinerStrategy;
    +			switch(operator.getCombineHint()) {
    +				case OPTIMIZER_CHOOSES:
    +					combinerStrategy = DriverStrategy.SORTED_PARTIAL_REDUCE;
    --- End diff --
    
    Merge the `OPTIMIZER_CHOOSES` and `SORT` cases?


---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message