You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Stephan Ewen (JIRA)" <ji...@apache.org> on 2014/08/29 19:15:54 UTC
[jira] [Commented] (FLINK-820) Support for unconnected data flows
[ https://issues.apache.org/jira/browse/FLINK-820?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14115507#comment-14115507 ]
Stephan Ewen commented on FLINK-820:
------------------------------------
This will be solved as a side effect of teh scheduler redesign as defined in [FLINK-1030].
> Support for unconnected data flows
> ----------------------------------
>
> Key: FLINK-820
> URL: https://issues.apache.org/jira/browse/FLINK-820
> Project: Flink
> Issue Type: Improvement
> Reporter: GitHub Import
> Labels: github-import
> Fix For: pre-apache
>
>
> Currently the compiler fails when given a plan with separate data flows.
> ```java
> throw new CompilerException("The given Pact program contains multiple disconnected data flows.");
> ```
> However, I do not see a good conceptual reason to not support such data flows. Are there any technical issues?
> To elaborate a bit on the issue: I have a higher level operator that may or may not produce separated data flows. In a current workaround I have to pipe the union of all data flows to a dummy sink to avoid such an exception.
> ---------------- Imported from GitHub ----------------
> Url: https://github.com/stratosphere/stratosphere/issues/820
> Created by: [AHeise|https://github.com/AHeise]
> Labels:
> Created at: Thu May 15 16:51:46 CEST 2014
> State: open
--
This message was sent by Atlassian JIRA
(v6.2#6252)