...

Lot of times, in Kafka Streams users want to send a record to more than one partition on the sink topic. Currently, if a user wants to replicate a message into N partitions, the only way of doing that is to replicate the message N times and then plug-in a new custom partitioner to write the message N times into N different partitions. To give a more concrete example, let's say there's a sink topic with 10 partitions and a user wants to send records only to even numbered partitions.

...

final int numPartitions = 10;

final KStream<String, String> inputStream = builder.stream("input-topic", Consumed.with(Serdes.String(), Serdes.String()));

for (int i = 0; i < numPartitions; i += 2) {
	final int evenNumberedPartition = i;
    inputStream.to("output-topic", Produced.with(Serdes.String(), Serdes.String(), (topic, key, val, numPartitions) -> evenNumberedPartition));
}

This seems a little cumbersome way to broadcast. Also, there seems to be no way of dropping a record within the partitionerAs can be seen, there's no implicit way of sending the records to multiple partitions. This KIP aims to make this process simpler in Kafka Streams. As a side note, this KIP also adds a provision to drop records using a custom partitioner.

Public Interfaces

The StreamPartitioner method would have a new method added called partitions()  and the current partition()  method would be marked as deprecated. The partitions()  method would be marked as  default within which it would invoke partition()  method and construct a singleton list out of it. Here's how the interface would look like now:

...