[
https://issues.apache.org/jira/browse/FLINK-2283?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14743545#comment-14743545
]
Aljoscha Krettek commented on FLINK-2283:
-----------------------------------------
I want to reiterate what I said earlier: I don't think it makes sense to keep
working on the non-windowed grouped operations. Everything that can be
expressed using one of them can also be expressed using grouped windows. There
are disadvantages to having separate implementation for them. We basically have
two implementations for very similar behaviour and have to ensure that both of
them work correctly. Once we manage to integrate the windowed state with a
state backend this also solves the state issue here. Also, the state in
non-windowed aggregation operations is infinite. This can make them very
dangerous to use in practice, in my opinion. If users are forced to manually
specify an infinite window when they want to have the behaviour they are at
least aware of what they are doing.
> Make grouped reduce/fold/aggregations stateful using Partitioned state
> ----------------------------------------------------------------------
>
> Key: FLINK-2283
> URL: https://issues.apache.org/jira/browse/FLINK-2283
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
> Affects Versions: 0.10
> Reporter: Gyula Fora
> Assignee: Márton Balassi
> Priority: Minor
>
> Currently the inner state of the grouped aggregations are not persisted as an
> operator state.
> These operators should be reimplemented to use the newly introduced
> partitioned state abstractions which will make them fault tolerant and
> scalable for the future.
> A suggested implementation would be to use a stateful mapper to implement the
> desired behaviour.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
