This module demonstrates the following:
- The usage of the Kafka Streams DSL, including
outerJoin()
between KStream and KStream,selectKey()
andpeek()
. - The usage of sliding time windows.
- Unit testing using Topology Test Driver.
In this module, records of type <String, KafkaPerson>
are streamed from two topics named PERSON_TOPIC
and PERSON_TOPIC_TWO
.
The following tasks are performed:
- Join the records on the last name within a 5-minute join window and a 1-minute grace period for delayed records.
- Build a new
KafkaJoinPersons
object that holds both persons. If no person is matched, a value holding the left or right person is still emitted as an outer join is performed. - Write the resulting
KafkaJoinPersons
objects to a new topic namedPERSON_OUTER_JOIN_STREAM_STREAM_TOPIC
.
To compile and run this demo, you will need the following:
- Java 21
- Maven
- Docker
To run the application manually, please follow the steps below:
- Start a Confluent Platform in a Docker environment.
- Produce records of type
<String, KafkaPerson>
to topics namedPERSON_TOPIC
andPERSON_TOPIC_TWO
. You can use the producer person to do this. - Start the Kafka Streams.
To run the application in Docker, please use the following command:
docker-compose up -d
This command will start the following services in Docker:
- 1 Kafka broker KRaft
- 1 Schema registry
- 1 Control Center
- 1 producer Person
- 1 Kafka Streams Outer Join Stream Stream