Skip to content

Consumed — Metadata for Consuming Records

Consumed<K, V> describes how to consume records in a topology in the High-Level KStream DSL for the following StreamsBuilder operators:

Consumed<K, V> is a NamedOperation.

Demo

import org.apache.kafka.common.serialization.Serdes
import org.apache.kafka.streams.kstream.Consumed
val consumed = Consumed.`with`(Serdes.Long, Serdes.String)

scala> :type consumed
org.apache.kafka.streams.kstream.Consumed[Long,String]

Creating Instance

Consumed takes the following to be created:

  • Serde<K> of keys (Apache Kafka)
  • Serde<V> of values (Apache Kafka)
  • TimestampExtractor
  • Reset Policy (Topology.AutoOffsetReset)
  • Processor Name

Consumed is created using the factories.

Creating Consumed

as

Consumed<K, V> as(
  String processorName)

with

Consumed<K, V> with(
  Serde<K> keySerde,
  Serde<V> valueSerde)
Consumed<K, V> with(
  Serde<K> keySerde,
  Serde<V> valueSerde,
  TimestampExtractor timestampExtractor,
  Topology.AutoOffsetReset resetPolicy)
Consumed<K, V> with(
  TimestampExtractor timestampExtractor)
Consumed<K, V> with(
  Topology.AutoOffsetReset resetPolicy)

Scala API

Scala API for Kafka Streams makes the optional Consumed metadata an implicit parameter in the StreamsBuilder API.

Moreover, ImplicitConversions object defines consumedFromSerde implicit method that creates a Consumed instance with the key and value Serde objects available in implicit scope.

And the last but not least, Scala API for Kafka Streams defines Consumed object with with factory methods that use implicit key and value Serde objects.

Back to top