You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chesnay Schepler (JIRA)" <ji...@apache.org> on 2015/02/06 13:58:34 UTC
[jira] [Commented] (FLINK-1418) Make 'print()' output on the client
command line, rather than on the task manager sysout
[ https://issues.apache.org/jira/browse/FLINK-1418?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14309085#comment-14309085 ]
Chesnay Schepler commented on FLINK-1418:
-----------------------------------------
is this not doable without breaking programs?
I'm thinking of something like this:
whenever a print sink is created, an accumulator is created that stores the output. the name of the accumulator is stored in a list in the ExEnv.
within env.execute(), after everything else ran (as in, before we return the JobExecutionResult), iterate through the accumulator entries stored and print the outputs the contents.
> Make 'print()' output on the client command line, rather than on the task manager sysout
> ----------------------------------------------------------------------------------------
>
> Key: FLINK-1418
> URL: https://issues.apache.org/jira/browse/FLINK-1418
> Project: Flink
> Issue Type: Improvement
> Components: Java API
> Affects Versions: 0.9
> Reporter: Stephan Ewen
>
> Right now, the {{print()}} command prints inside the data sinks where the code runs. It should pull data back to the client and print it there.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)