kafka-streams-aggregate-hopping-window

Kafka Streams Aggregate Hopping Window

This module demonstrates the following:

• The use of the Kafka Streams DSL, including aggregate(), windowedBy().advanceBy(), groupByKey(), selectKey(), toStream() and peek().
• Unit testing using Topology Test Driver.

In this module, records of type <String, KafkaUser> are streamed from a topic named USER_TOPIC. The following tasks are performed:

1. Group the stream by last name using groupByKey() operation.
2. Apply an aggregator that combines each KafkaUser record with the same last name into a KafkaUserGroup object and aggregates the first names by last name.
3. The aggregations are performed using a 5-minute time window with a 2-minute hop, and a 1-minute grace period for delayed records.
4. Write the resulting records to a new topic named USER_AGGREGATE_HOPPING_WINDOW_TOPIC.

The output records will be in the following format:

{"firstNameByLastName":{"Last name 1":{"First name 1", "First name 2", "First name 3"}}}
{"firstNameByLastName":{"Last name 2":{"First name 4", "First name 5", "First name 6"}}}
{"firstNameByLastName":{"Last name 3":{"First name 7", "First name 8", "First name 9"}}}

Prerequisites

To compile and run this demo, you will need the following:

• Java 21
• Maven
• Docker

Running the Application

To run the application manually:

• Start a Confluent Platform in a Docker environment.
• Produce records of type <String, KafkaUser> to a topic named USER_TOPIC. You can use the producer user to do this.
• Start the Kafka Streams application.

To run the application in Docker, use the following command:

docker-compose up -d

This command will start the following services in Docker:

• 1 Kafka broker (KRaft mode)
• 1 Schema registry
• 1 Control Center
• 1 producer User
• 1 Kafka Streams Aggregate Hopping Window