Kafka Streams App - count and sum aggregate

Sam Shiles :

I'm trying to create a KTable from a KGroupedStream to store the sum of the value for each key.

 final StreamsBuilder builder = new StreamsBuilder();
 final KTable<String, Long> sum = builder.stream("streams-plaintext-input")
            .groupByKey()
            .aggregate(new Initializer<Long>() {
                @Override
                public Long apply() {
                    return Long.MIN_VALUE;
                }
            }, new Aggregator<String, Long, Long>() {
                @Override
                public Long apply(final String key, final Long value,final Long aggregate) {
                    aggregate += value;
                    return aggregate;
                }
            }, Materialized.<String, Long, KeyValueStore<Byte, byte[]>>as("counts-store"));

but am getting the error:

The method aggregate(Initializer<VR>, Aggregator<? super Object,? super Object,VR>, Materialized<Object,VR,KeyValueStore<Bytes,byte[]>>) in the type KGroupedStream<Object,Object> is not applicable for the arguments (new Initializer<Long>(){}, new Aggregator<String,Long,Long>(){}, Materialized<String,Long,KeyValueStore<Byte,byte[]>>)

All the examples I've seen pass in a Serde as the third argument but I have tried this and get a very similar error (I think this might be from an older version as it does not match with the signature of the current implementation?):

final StreamsBuilder builder = new StreamsBuilder();
    final KTable<String, Long> sum = builder.stream("streams-plaintext-input")
            .groupByKey()
            .aggregate(new Initializer<Long>() {
                @Override
                public Long apply() {
                    return Long.MIN_VALUE;
                }
            }, new Aggregator<String, Long, Long>() {
                @Override
                public Long apply(final String key, final Long value,final Long aggregate) {
                    aggregate += value;
                    return aggregate;
                }
            }, Serdes.Long());

Error:

The method aggregate(Initializer<VR>, Aggregator<? super Object,? super Object,VR>, Materialized<Object,VR,KeyValueStore<Bytes,byte[]>>) in the type KGroupedStream<Object,Object> is not applicable for the arguments (new Initializer<Long>(){}, new Aggregator<String,Long,Long>(){}, Serde<Long>)

What am I doing wrong?

Using Kafka version: 2.1.0

Bartosz Wardziński :

There are few issues in your code:

1. For Materialized.as instead java.lang.Byte you should pass org.apache.kafka.common.utils.Bytes
2. You shouldn't modify final variable: aggregate += value;
3. You have to add types of key and value to StreamsBuilder::stream call (builder.<String, Long>stream("streams-plaintext-input"))

After modification it should looks more or less as follow:

KTable<String, Long> sum = builder.<String, Long>stream("streams-plaintext-input")
        .groupByKey()
        .aggregate(new Initializer<Long>() {
            @Override
            public Long apply() {
                return Long.MIN_VALUE;
            }
        }, new Aggregator<String, Long, Long>() {
            @Override
            public Long apply(final String key, final Long value,final Long aggregate) {
                return aggregate + value;
            }
        }, Materialized.<String, Long, KeyValueStore<Bytes, byte[]>>as("counts-store"));