The comment from Kasier Chen on the patch is correct:
1. Objects.equals() depends on the typed class object.equal() function,
which may not be safe: for example users can override the equals function
to only compare a subset of all the fields, whereas for Kafka Streams we
need these two objects
Hi Guozhang,
I've been working around this issue by dropping down to the Processor
API, but, I was hoping you might be able to point out if there is a
flaw is in this proposed change:
https://github.com/apache/kafka/compare/trunk...mfenniak:suppress-duplicate-repartition-output
This
Hi Mathieu,
If you are only interested in the aggregate result "snapshot" but not its
change stream (note that KTable itself is not actually a "table" as in
RDBMS, but still a stream), you can try to use the queryable state feature
that is available in trunk, which will be available in 0.10.1.0
The problem is that Kafka Streams need to repartition the streams based on
the groupBy keys when doing aggregations. For your case, the original
stream may be read from a topic that is partitioned on "K", and you need to
first repartition on "category" on an intermediate topic before the
Hello again, kafka-users,
When I aggregate a KTable, a future input that updates a KTable's
value for a specific key causes the aggregate's subtractor to be
invoked, and then its adder. This part is great, completely
as-expected.
But what I didn't expect is that the intermediate result of the