[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14646600#comment-14646600 ] ASF GitHub Bot commented on FLINK-2419: --- Github user asfgit closed the pull request at: https://github.com/apache/flink/pull/947 > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14646592#comment-14646592 ] ASF GitHub Bot commented on FLINK-2419: --- Github user senorcarbone commented on the pull request: https://github.com/apache/flink/pull/947#issuecomment-126054868 :+1: from me. looks that it does the job > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14645779#comment-14645779 ] ASF GitHub Bot commented on FLINK-2419: --- Github user gyfora commented on the pull request: https://github.com/apache/flink/pull/947#issuecomment-125894533 If you give a +1 @rmetzger I will merge this > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644962#comment-14644962 ] ASF GitHub Bot commented on FLINK-2419: --- GitHub user gyfora opened a pull request: https://github.com/apache/flink/pull/947 [FLINK-2419] Add test for sinks after keyBy and groupBy Adds a missing test for FLINK-2419 You can merge this pull request into a Git repository by running: $ git pull https://github.com/gyfora/flink sink-hotfix Alternatively you can review and apply these changes as the patch at: https://github.com/apache/flink/pull/947.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #947 > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644674#comment-14644674 ] Gyula Fora commented on FLINK-2419: --- I added an issue for this: https://issues.apache.org/jira/browse/FLINK-2423 > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644666#comment-14644666 ] Robert Metzger commented on FLINK-2419: --- I see that its tricky to implement a test for this, but its one of the main features of Flink, so its worth the effort (in my opinion) I don't have time to work on this right now. I agree that opening a second issue is a good idea, because the commit is actually fixing two bugs. With two separate issues, we can track the resolution of the issues easier. > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[
https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644510#comment-14644510
]
Robert Metzger commented on FLINK-2419:
---
Okay, I suspect its this one: https://github.com/apache/flink/pull/937
I'll still re-open the issue until we have a test that ensures that the
{{notifyCheckpointComplete}} is called exactly once.
> DataStream sinks lose key information
> -
>
> Key: FLINK-2419
> URL: https://issues.apache.org/jira/browse/FLINK-2419
> Project: Flink
> Issue Type: Bug
> Components: Streaming
>Reporter: Gyula Fora
>Assignee: Gyula Fora
>Priority: Blocker
> Fix For: 0.10
>
>
> If the user applies an addSink method after keyBy, the sink will not have the
> information of the key as the transform method is bypassed.
> This makes it impossible to use partitioned states with sinks.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644526#comment-14644526 ] Gyula Fora commented on FLINK-2419: --- Alright, it's going to be a tricky test due to the lack of guarantees we provide for the notifications. I can't work on this right now, feel free to give it a shot if you want. I will actually open a separate Jira for this when i get home. > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644503#comment-14644503 ] Gyula Fora commented on FLINK-2419: --- I added a test in my other PR for the sinks. > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (FLINK-2419) DataStream sinks lose key information
[ https://issues.apache.org/jira/browse/FLINK-2419?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14644487#comment-14644487 ] Robert Metzger commented on FLINK-2419: --- Do you agree to reopen the issue until a test case has been implemented for this? > DataStream sinks lose key information > - > > Key: FLINK-2419 > URL: https://issues.apache.org/jira/browse/FLINK-2419 > Project: Flink > Issue Type: Bug > Components: Streaming >Reporter: Gyula Fora >Assignee: Gyula Fora >Priority: Blocker > Fix For: 0.10 > > > If the user applies an addSink method after keyBy, the sink will not have the > information of the key as the transform method is bypassed. > This makes it impossible to use partitioned states with sinks. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
