You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Timo Walther (Jira)" <ji...@apache.org> on 2020/10/06 08:36:00 UTC
[jira] [Commented] (FLINK-19508) Add collect() operation on
DataStream
[ https://issues.apache.org/jira/browse/FLINK-19508?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17208567#comment-17208567 ]
Timo Walther commented on FLINK-19508:
--------------------------------------
+1 but we should make it clear that this method performs an execution and `env.execute()` would throw an exception. Thus, I would propose to name it `DataStream.executeAndCollect()`, `DataStream.executeAndPrint()` or similar.
> Add collect() operation on DataStream
> -------------------------------------
>
> Key: FLINK-19508
> URL: https://issues.apache.org/jira/browse/FLINK-19508
> Project: Flink
> Issue Type: Improvement
> Components: API / DataStream
> Reporter: Aljoscha Krettek
> Priority: Major
>
> With the recent changes/additions to {{DataStreamUtils.collect()}} that make it more robust by using the regular REST client to fetch results from operators it might make sense to add a {{collect()}} operation right on {{DataStream}}.
> This operation is still not meant for big data volumes but I think it can be useful for debugging and fetching small amounts of messages to the client. When we do this, we can also think about changing {{print()}} to print on the client instead of to the {{TaskManager}} stdout. I think the current behaviour of this operation is mostly confusing for users.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)