You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Chiwan Park (JIRA)" <ji...@apache.org> on 2015/06/02 07:57:17 UTC
[jira] [Commented] (FLINK-2069) writeAsCSV function in DataStream
Scala API creates no file
[ https://issues.apache.org/jira/browse/FLINK-2069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14568579#comment-14568579 ]
Chiwan Park commented on FLINK-2069:
------------------------------------
Hi. I am concerned about decreasing performance. Currently, {{CsvOutputFormat}} uses {{BufferedOutputStream}} to increase writing performance. In streaming job, Calling {{flush()}} is reasonable to create output immediately. But in batch job, the calling {{flush()}} for every output record cancels the benefit of buffered output.
What do you think?
> writeAsCSV function in DataStream Scala API creates no file
> -----------------------------------------------------------
>
> Key: FLINK-2069
> URL: https://issues.apache.org/jira/browse/FLINK-2069
> Project: Flink
> Issue Type: Bug
> Components: Streaming
> Reporter: Faye Beligianni
> Priority: Blocker
> Labels: Streaming
> Fix For: 0.9
>
>
> When the {{writeAsCSV}} function is used in the DataStream Scala API, no file is created in the specified path.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)