Christopher Meier created KAFKA-14268:
-----------------------------------------
Summary: When aggregating on a KGroupedStream allow an
InitializerWithKey
Key: KAFKA-14268
URL: https://issues.apache.org/jira/browse/KAFKA-14268
Project: Kafka
Issue Type: Improvement
Components: streams
Reporter: Christopher Meier
Sometimes when aggregating on a KGroupedStream I would like to have an initial
value which change depending on the key.
A workarround is to have a dummy value in the initializer which can be checked
for and replaced in the aggregator, but that implies some performance loss
caused by the check.
My proposal is to add yet another overload for {{aggregate}} which has instead
of an {{Initializer}} an {{{}InitializerWithKey{}}}. The {{apply}} method of
{{InitializerWithKey}} would have the key as input and the initial value as
output.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
