I agree, but IMHO it's exposing some unintuitive behavior with respect to 
creating multiple repartition topics.  And yes with optimizations turned on 
users will be able to write code in this form. 

But as it stands now, by explicitly naming a repartition I don't see how we can 
re-use a single `KGroupedStream` instance, as before we relied on the 
auto-generated names to handle the creation of multiple repartition topics.  

>Maybe we can merge both repartition topics into one for this case, too?

I'm not sure I follow, do you mean in this case we do an automatic optimization 
and merge repartition topics "in-line"?   If so, I'm inclined to say yes we 
can, but I'm thinking this may be done best in a follow-on PR.

WDYT?

[ Full content available at: https://github.com/apache/kafka/pull/5709 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to