[
https://issues.apache.org/jira/browse/FLINK-18840?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17173002#comment-17173002
]
Nico Kruber commented on FLINK-18840:
-------------------------------------
Thinking about this wholistically and about the users' view on this, it is a
bit unclear why you have to call {{StreamExecutionEnvironment#execute()}} in
the first place whenever you convert to {{DataStream}}. Ideally, either
{{execute}} call ({{Table#executeInsert()}}, {{StatementSet#execute()}}, or
{{StreamExecutionEnvironment#execute()}}) should always execute everything if
you allow mixing these APIs. That, I guess, would be the ideal solution.
>From the code's perspective this would probably need a tighter integration...
>At least the conversion to {{DataStream}}, though, could register this stream
>in the TableAPI somehow so that it will be taken into account during its
>execute calls. Getting {{StreamExecutionEnvironment#execute()}} to run the
>TableAPI parts is probably more complex / would couple the two more than
>(probably) wanted.
> Support StatementSet with DataStream API
> ----------------------------------------
>
> Key: FLINK-18840
> URL: https://issues.apache.org/jira/browse/FLINK-18840
> Project: Flink
> Issue Type: New Feature
> Components: Table SQL / API
> Reporter: Timo Walther
> Priority: Major
>
> Currently, users of the {{StreamTableEnvironment}} cannot not translate a
> {{StatementSet}} to DataStream API.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
