0

I have this kstream topology:

KStream<String, Data> eventStream = streamsBuilder.stream("event",
        Consumed.with(Serdes.String(), eventSerde);

KTable<String, Result> resultKTable = streamsBuilder.table("result",
        Consumed.with(Serdes.String(), resultSerde));

eventStream.leftJoin(resultKTable, new Joiner())
        .to("result", Produced.with(Serdes.String(), resultSerde));

When I sent tombstone record in "event" topic, it gives me this error

Skipping record due to null join key or value. key=[3428642] value=[null] topic=[event] partition=[0] offset=[21]

EDIT
I did the following workaround to handle tombstone records:

eventStream.filter((k,v) -> v != null)
           .leftJoin(resultKTable, new Joiner())
           .to("result", Produced.with(Serdes.String(), resultSerde));
eventStream.filter((k,v) -> v == null).to("result");
cppcoder
  • 22,227
  • 6
  • 56
  • 81

1 Answers1

1

It is not an error, but warning.

As written in Kafka Streams documentation for KStream-KTable left join:

Input records for the stream with a null key or a null value are ignored and do not trigger the join

So, this is expected scenario.

Side note: There is written blogpost at luppeng.wordpress.com about this and possible workaround, which might come handy.

hradecek
  • 2,455
  • 2
  • 21
  • 30
  • Got it. I did another workaround to handle tombstones. Is that a right approach? Updated the question – cppcoder Mar 17 '23 at 05:21
  • Looks good to me (you might also use split or merge streams before sending to sink "result" topic" - but in this case there are no real benefits) – hradecek Mar 17 '23 at 11:07