[
https://issues.apache.org/jira/browse/FLINK-3477?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15369666#comment-15369666
]
ASF GitHub Bot commented on FLINK-3477:
---------------------------------------
Github user ggevay commented on a diff in the pull request:
https://github.com/apache/flink/pull/1517#discussion_r70182792
--- 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 --
I included the `OPTIMIZER_CHOOSES` case because `JoinHint` and `CrossHint`
also have this case. I guess it is there because we might modify what the
default is later.
> Add hash-based combine strategy for ReduceFunction
> --------------------------------------------------
>
> Key: FLINK-3477
> URL: https://issues.apache.org/jira/browse/FLINK-3477
> Project: Flink
> Issue Type: Sub-task
> Components: Local Runtime
> Reporter: Fabian Hueske
> Assignee: Gabor Gevay
>
> This issue is about adding a hash-based combine strategy for ReduceFunctions.
> The interface of the {{reduce()}} method is as follows:
> {code}
> public T reduce(T v1, T v2)
> {code}
> Input type and output type are identical and the function returns only a
> single value. A Reduce function is incrementally applied to compute a final
> aggregated value. This allows to hold the preaggregated value in a hash-table
> and update it with each function call.
> The hash-based strategy requires special implementation of an in-memory hash
> table. The hash table should support in place updates of elements (if the
> updated value has the same size as the new value) but also appending updates
> with invalidation of the old value (if the binary length of the new value
> differs). The hash table needs to be able to evict and emit all elements if
> it runs out-of-memory.
> We should also add {{HASH}} and {{SORT}} compiler hints to
> {{DataSet.reduce()}} and {{Grouping.reduce()}} to allow users to pick the
> execution strategy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
