[
https://issues.apache.org/jira/browse/KAFKA-7445?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16633103#comment-16633103
]
Matthias J. Sax commented on KAFKA-7445:
----------------------------------------
I guess, you can write the following code to get what you need:
{noformat}
KStream stream = ...
KStream stream1 = stream.filter(...);
KStream stream2 = stream.filter(...);
KStream stream3 = stream.filter(...);{noformat}
This code applies each `Predicate` to all records. Reusing the same `KStream`
variable (ie, `stream`) multiple times is a logical broadcast, ie, each record
will be passed into each `filter()`.
It's up to you if you think it would be helpful to have a single operator for
this. The question is, how this operator could be named?
> Branch one Stream in multiple Streams
> -------------------------------------
>
> Key: KAFKA-7445
> URL: https://issues.apache.org/jira/browse/KAFKA-7445
> Project: Kafka
> Issue Type: New Feature
> Components: streams
> Reporter: Dennis Reiter
> Priority: Minor
>
> Hi,
> I need to branch/split KStreams in multiple independent KStreams. I thought,
> {{org.apache.kafka.streams.kstream.internals.KStreamImpl#branch}} is the
> right one but in fact, its designed for another purpose.
> In contrast to {{branch}} I need to assign the record to *all* matching
> streams, not only one stream.
> Speaking in code
> ({{org.apache.kafka.streams.kstream.internals.KStreamBranch}}):
> {code:java}
> if (predicates[i].test(key, value)) {
> // use forward with childIndex here
> // and pipe the record to multiple streams
> context().forward(key, value, i);
> }
> {code}
> My question: is this still possible with features already included in
> Streams? Or shall I propose a change?
> Thanks in advance
> Dennis
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
