[
https://issues.apache.org/jira/browse/FLINK-1991?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14534281#comment-14534281
]
Fabian Hueske commented on FLINK-1991:
--------------------------------------
I know, but `Pojo` or `Row` are not good solutions either.
One reason to use the Table API is to reduce UDF bloat. A lot of that advantage
is gone if I have to specify a Pojo object (which is even more bloat than
defining a Tuple return type) to continue operating on the data.
> Return Table as DataSet<Tuple>
> ------------------------------
>
> Key: FLINK-1991
> URL: https://issues.apache.org/jira/browse/FLINK-1991
> Project: Flink
> Issue Type: Improvement
> Components: Table API
> Affects Versions: 0.9
> Reporter: Fabian Hueske
>
> WIth a {{TableEnvironment}} a {{Table}} can be converted into a Pojo or Row
> data set. Both types have their issues. A Pojo needs to be explicitly defined
> and a Row cannot be easily processed by a user-function.
> It would be good to have a method to convert a {{Table}} into a
> {{DataSet<Tuple>}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
