[
https://issues.apache.org/jira/browse/FLINK-4374?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15417617#comment-15417617
]
Stephan Ewen commented on FLINK-4374:
-------------------------------------
True, but the test should not try a Tuple2<X, Date> if tuples do not handle
nulls. It's bound to produce something unpredictable.
> GroupReduce Broken for null Date
> --------------------------------
>
> Key: FLINK-4374
> URL: https://issues.apache.org/jira/browse/FLINK-4374
> Project: Flink
> Issue Type: Bug
> Components: DataSet API
> Reporter: Stefan Richter
> Assignee: Timo Walther
>
> The GroupReduceITCase has an error that allows a problem with {{null}} Dates
> to go uncovered:
> If I set the parallelism to 1 in {{testDateNullException()}} and all keys
> actually end up on the same operator, then there is a problem in the
> de/serialization.
> It seems that {{null}} values are somehow skipped by the serialization
> process (e.g. maybe no {{null}} indicator is written), which leads to wrong
> deserializations.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
