[
https://issues.apache.org/jira/browse/FLINK-5727?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15855627#comment-15855627
]
ASF GitHub Bot commented on FLINK-5727:
---------------------------------------
Github user KurtYoung commented on the issue:
https://github.com/apache/flink/pull/3281
Hi @fhueske , thanks for your reviewing and comments, it makes sense to me.
The motivation behind this refactoring is we want to eliminate most differences
between batch and stream execution environment, and use `TableEnvironment` only
when they writing programs.
How about we still keep `registerTableSource` in `TableEnvironment` and
leave the implementation to the specific env, and do some type validation
before registration.
> Unify some API of batch and stream TableEnvironment
> ---------------------------------------------------
>
> Key: FLINK-5727
> URL: https://issues.apache.org/jira/browse/FLINK-5727
> Project: Flink
> Issue Type: Improvement
> Components: Table API & SQL
> Reporter: Kurt Young
> Assignee: Kurt Young
>
> Some API's implementation of current {{BatchTableEnvironment}} and
> {{StreamTableEnvironment}} is identical, like {{sql}}. And other API like
> {{registerTableSource}} and {{scan}} can be unified to the base class. (we
> can change {{ingest}} from {{StreamTableEnvironment}} to {{scan}} also)
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
