[
https://issues.apache.org/jira/browse/FLINK-6388?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15988313#comment-15988313
]
ASF GitHub Bot commented on FLINK-6388:
---------------------------------------
Github user huawei-flink commented on the issue:
https://github.com/apache/flink/pull/3783
@rtudoran @fhueske the first implementation I made was with the state in
the ProcessFunction without code generated aggregation function. Second, I
pushed a branch with the state in the process function using the code generated
process function. Then, third I moved the state within the code generated
function.
It is not clear to me why the state cannot be within the code generated
function. Could you please clarify so that we can understand whether it is
worth working around it. This feature is quite important for us.
Anyway, you could have a look at the branch that uses the state in the
process function and uses the code generated aggregation functions. Basically,
rather than generate one code generated function for all the aggregations, I
create one class for each, and then I call the corresponding accumulate/retract
using the distinct logic when marked in the process function.
> Add support for DISTINCT into Code Generated Aggregations
> ---------------------------------------------------------
>
> Key: FLINK-6388
> URL: https://issues.apache.org/jira/browse/FLINK-6388
> Project: Flink
> Issue Type: Sub-task
> Components: DataStream API
> Affects Versions: 1.3.0
> Reporter: Stefano Bortoli
> Assignee: Stefano Bortoli
> Fix For: 1.3.0
>
>
> We should support DISTINCT in Code Generated aggrgation functions.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
