Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/1255#discussion_r43987736
--- Diff:
flink-runtime/src/main/java/org/apache/flink/runtime/operators/shipping/OutputEmitter.java
---
@@ -135,7 +116,7 @@ public OutputEmitter(ShipStrategyType strategy,
TypeComparator<T> comparator, Pa
case PARTITION_CUSTOM:
return customPartition(record.getInstance(),
numberOfChannels);
case PARTITION_RANGE:
- return rangePartition(record.getInstance(),
numberOfChannels);
+ return rangePartition((Tuple2<Integer,
?>)record.getInstance(), numberOfChannels);
--- End diff --
Yes, the current version of `DataDistribution` was supported by a previous
version of the API (even pre-Flink). With the generalized data model,
`DataDistribution` should be backed by a `Object[][]` instead of `Key<?>[]` and
the `TypeComparator`s should be used to compare the individual key fields.
I think if we change DataDistribution now to operate on `Object[][]`, the
follow up task would become much easier. Also if you plan to add a
`DataDistribution` later, wouldn't it make more sense to update it now, instead
of removing it first and adding an updated version later again?
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
