[ 
https://issues.apache.org/jira/browse/FLINK-1034?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14077734#comment-14077734
 ] 

Kostas Tzoumas commented on FLINK-1034:
---------------------------------------

Overloading the DataSet.reduce method leads to an ugly case: if Java 8 lambdas 
will be supported, the Java compiler cannot distinguish between a 
GroupReduceFunction and a ReduceFunction in the form of lambdas without the 
user explicitly providing types.

This is uglier than having two top-level methods.

Therefore, I propose to leave things as are.

> Inconsistent naming of DataSet reduceGroup and GroupReduceFunction
> ------------------------------------------------------------------
>
>                 Key: FLINK-1034
>                 URL: https://issues.apache.org/jira/browse/FLINK-1034
>             Project: Flink
>          Issue Type: Improvement
>          Components: Java API
>            Reporter: Ufuk Celebi
>            Priority: Trivial
>              Labels: breaking-api
>
> I don't know if it is on purpose or not, but I found it to be inconsistent 
> and sometimes also confusing:
> Why not
> {code:java}
> input.groupReduce(new GroupReduceFunction() {...})
> {code}
> instead of the current
> {code:java}
> input.reduceGroup(new GroupReduceFunction() {...})
> {code}
> I think it's the same line of thinking, which results in overriding a 
> {{reduce}} method in {{GroupReduceFunction}} instead of a {{groupReduce}} 
> method.
> Is there a nice way to unify this?



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to