How to merge many streams into one stream

Question:

If you have many Kafka topics with events, how do you merge them all into a single topic?

Example use case:

Suppose that you have a set of Kafka topics representing songs of a particular genre being played. You might have a topic for rock songs, another for classical songs, and so forth. In this tutorial, we'll write a program that merges all of the song play events into a single topic. Related pattern: Event Stream Merger.

Hands-on code example:

Confluent Cloud ksqlDB Kafka Streams Basic Kafka

New to Confluent Cloud? Get started here.

Short Answer

The input streams are combined using the merge function, which creates a new stream that represents all of the events of its inputs. The merged stream is forwarded to a combined topic via the to method, which accepts the topic as a parameter.

KStream<String, SongEvent> rockSongs = builder.stream(rockTopic);
KStream<String, SongEvent> classicalSongs = builder.stream(classicalTopic);
KStream<String, SongEvent> allSongs = rockSongs.merge(classicalSongs);

allSongs.to(allGenresTopic);

Run it

Initialize the project

To get started, make a new directory anywhere you’d like for this project:

mkdir merge-streams && cd merge-streams

Next, create a directory for configuration data:

mkdir configuration

Provision your Kafka cluster

This tutorial requires access to an Apache Kafka cluster, and the quickest way to get started free is on Confluent Cloud, which provides Kafka as a fully managed service.

Take me to Confluent Cloud

After you log in to Confluent Cloud, click Environments in the lefthand navigation, click on Add cloud environment, and name the environment learn-kafka. Using a new environment keeps your learning resources separate from your other Confluent Cloud resources.
From the Billing & payment section in the menu, apply the promo code CC100KTS to receive an additional $100 free usage on Confluent Cloud (details). To avoid having to enter a credit card, add an additional promo code CONFLUENTDEV1. With this promo code, you will not have to enter a credit card for 30 days or until your credits run out.
Click on LEARN and follow the instructions to launch a Kafka cluster and enable Schema Registry.

Write the cluster information into a local file

From the Confluent Cloud Console, navigate to your Kafka cluster and then select Clients in the lefthand navigation. From the Clients view, create a new client and click Java to get the connection information customized to your cluster.

Create new credentials for your Kafka cluster and Schema Registry, writing in appropriate descriptions so that the keys are easy to find and delete later. The Confluent Cloud Console will show a configuration similar to below with your new credentials automatically populated (make sure Show API keys is checked). Copy and paste it into a configuration/ccloud.properties file on your machine.

# Required connection configs for Kafka producer, consumer, and admin
bootstrap.servers={{ BOOTSTRAP_SERVERS }}
security.protocol=SASL_SSL
sasl.jaas.config=org.apache.kafka.common.security.plain.PlainLoginModule required username='{{ CLUSTER_API_KEY }}' password='{{ CLUSTER_API_SECRET }}';
sasl.mechanism=PLAIN
# Required for correctness in Apache Kafka clients prior to 2.6
client.dns.lookup=use_all_dns_ips

# Best practice for Kafka producer to prevent data loss
acks=all

# Required connection configs for Confluent Cloud Schema Registry
schema.registry.url={{ SR_URL }}
basic.auth.credentials.source=USER_INFO
basic.auth.user.info={{ SR_API_KEY }}:{{ SR_API_SECRET }}

Do not directly copy and paste the above configuration. You must copy it from the Confluent Cloud Console so that it includes your Confluent Cloud information and credentials.

Download and set up the Confluent CLI

This tutorial has some steps for Kafka topic management and producing and consuming events, for which you can use the Confluent Cloud Console or the Confluent CLI. Follow the instructions here to install the Confluent CLI, and then follow these steps connect the CLI to your Confluent Cloud cluster.

Configure the project

Create the following Gradle build file, named build.gradle for the project:

buildscript {
    repositories {
        mavenCentral()
    }
    dependencies {
        classpath "gradle.plugin.com.github.jengelman.gradle.plugins:shadow:7.0.0"
    }
}

plugins {
    id "java"
    id "com.github.davidmc24.gradle.plugin.avro" version "1.7.0"
}

sourceCompatibility = JavaVersion.VERSION_17
targetCompatibility = JavaVersion.VERSION_17
version = "0.0.1"

repositories {
    mavenCentral()

    maven {
        url "https://packages.confluent.io/maven"
    }
}

apply plugin: "com.github.johnrengelman.shadow"

dependencies {
    implementation "org.apache.avro:avro:1.11.1"
    implementation "org.slf4j:slf4j-simple:2.0.7"
    implementation 'org.apache.kafka:kafka-streams:3.4.0'
    implementation ('org.apache.kafka:kafka-clients') {
       version {
           strictly '3.4.0'
        }
      }
    implementation "io.confluent:kafka-streams-avro-serde:7.3.0"
    testImplementation "org.apache.kafka:kafka-streams-test-utils:3.4.0"
    testImplementation "junit:junit:4.13.2"
}

test {
    testLogging {
        outputs.upToDateWhen { false }
        showStandardStreams = true
        exceptionFormat = "full"
    }
}

jar {
  manifest {
    attributes(
      "Class-Path": configurations.compileClasspath.collect { it.getName() }.join(" "),
      "Main-Class": "io.confluent.developer.MergeStreams"
    )
  }
}

shadowJar {
    archiveBaseName = "kstreams-merge-standalone"
    archiveClassifier = ''
}

And be sure to run the following command to obtain the Gradle wrapper:

gradle wrapper

Then create a development configuration file at configuration/dev.properties:

application.id=merging-app
replication.factor=3

input.rock.topic.name=rock-song-events
input.rock.topic.partitions=6
input.rock.topic.replication.factor=3

input.classical.topic.name=classical-song-events
input.classical.topic.partitions=6
input.classical.topic.replication.factor=3

output.topic.name=all-song-events
output.topic.partitions=6
output.topic.replication.factor=3

Update the properties file with Confluent Cloud information

Using the command below, append the contents of configuration/ccloud.properties (with your Confluent Cloud configuration) to configuration/dev.properties (with the application properties).

cat configuration/ccloud.properties >> configuration/dev.properties

Create a schema for the events

Create a directory for the schemas that represent the events in the stream:

mkdir -p src/main/avro

Then create the following Avro schema file at src/main/avro/song_event.avsc for the events representing a song being played:

{
  "namespace": "io.confluent.developer.avro",
  "type": "record",
  "name": "SongEvent",
  "fields": [
    {"name": "artist", "type": "string"},
    {"name": "title", "type": "string"}
  ]
}

Because we will use this Avro schema in our Java code, we’ll need to compile it. Run the following:

./gradlew build

Create the Kafka Streams topology

Create a directory for the Java files in this project:

mkdir -p src/main/java/io/confluent/developer

Then create the following file at src/main/java/io/confluent/developer/MergeStreams.java. Notice the buildTopology method, which uses the Kafka Streams DSL. A stream is opened up for each input topic. The input streams are then combined using the merge function, which creates a new stream that represents all of the events of its inputs. Note that you can chain merge to combine as many streams as needed. The merged stream is then connected to the to method, which the name of a Kafka topic to send the events to.

package io.confluent.developer;

import org.apache.kafka.clients.admin.AdminClient;
import org.apache.kafka.clients.admin.NewTopic;
import org.apache.kafka.common.serialization.Serdes;
import org.apache.kafka.streams.KafkaStreams;
import org.apache.kafka.streams.StreamsBuilder;
import org.apache.kafka.streams.StreamsConfig;
import org.apache.kafka.streams.Topology;
import org.apache.kafka.streams.kstream.KStream;

import java.io.FileInputStream;
import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.time.Duration;
import java.util.Properties;
import java.util.concurrent.CountDownLatch;

import io.confluent.developer.avro.SongEvent;
import io.confluent.kafka.streams.serdes.avro.SpecificAvroSerde;

import static io.confluent.kafka.serializers.AbstractKafkaSchemaSerDeConfig.SCHEMA_REGISTRY_URL_CONFIG;

public class MergeStreams {

    public Topology buildTopology(Properties allProps) {
        final StreamsBuilder builder = new StreamsBuilder();

        final String rockTopic = allProps.getProperty("input.rock.topic.name");
        final String classicalTopic = allProps.getProperty("input.classical.topic.name");
        final String allGenresTopic = allProps.getProperty("output.topic.name");

        KStream<String, SongEvent> rockSongs = builder.stream(rockTopic);
        KStream<String, SongEvent> classicalSongs = builder.stream(classicalTopic);
        KStream<String, SongEvent> allSongs = rockSongs.merge(classicalSongs);

        allSongs.to(allGenresTopic);
        return builder.build();
    }

    public void createTopics(Properties allProps) {
        AdminClient client = AdminClient.create(allProps);

        List<NewTopic> topics = new ArrayList<>();

        topics.add(new NewTopic(
                allProps.getProperty("input.rock.topic.name"),
                Integer.parseInt(allProps.getProperty("input.rock.topic.partitions")),
                Short.parseShort(allProps.getProperty("input.rock.topic.replication.factor"))));

        topics.add(new NewTopic(
                allProps.getProperty("input.classical.topic.name"),
                Integer.parseInt(allProps.getProperty("input.classical.topic.partitions")),
                Short.parseShort(allProps.getProperty("input.classical.topic.replication.factor"))));

        topics.add(new NewTopic(
                allProps.getProperty("output.topic.name"),
                Integer.parseInt(allProps.getProperty("output.topic.partitions")),
                Short.parseShort(allProps.getProperty("output.topic.replication.factor"))));

        client.createTopics(topics);
        client.close();
    }

    public Properties loadEnvProperties(String fileName) throws IOException {
        Properties allProps = new Properties();
        FileInputStream input = new FileInputStream(fileName);
        allProps.load(input);
        input.close();

        return allProps;
    }

    public static void main(String[] args) throws Exception {
        if (args.length < 1) {
            throw new IllegalArgumentException("This program takes one argument: the path to an environment configuration file.");
        }

        MergeStreams ms = new MergeStreams();
        Properties allProps = ms.loadEnvProperties(args[0]);
        allProps.put(StreamsConfig.DEFAULT_KEY_SERDE_CLASS_CONFIG, Serdes.String().getClass());
        allProps.put(StreamsConfig.DEFAULT_VALUE_SERDE_CLASS_CONFIG, SpecificAvroSerde.class);
        allProps.put(SCHEMA_REGISTRY_URL_CONFIG, allProps.getProperty("schema.registry.url"));
        Topology topology = ms.buildTopology(allProps);

        ms.createTopics(allProps);

        final KafkaStreams streams = new KafkaStreams(topology, allProps);
        final CountDownLatch latch = new CountDownLatch(1);

        // Attach shutdown handler to catch Control-C.
        Runtime.getRuntime().addShutdownHook(new Thread("streams-shutdown-hook") {
            @Override
            public void run() {
                streams.close(Duration.ofSeconds(5));
                latch.countDown();
            }
        });

        try {
            streams.start();
            latch.await();
        } catch (Throwable e) {
            System.exit(1);
        }
        System.exit(0);
    }
}

Note when using the merge operator the keys and values of the two KStream objects you’re merging must be of the same type. If you have 2 KStream instances with different key and/or value types, you’ll have to use the KStream.map (or KStream.mapValues) operation first to get the types to line-up before merging.

Compile and run the Kafka Streams program

In your terminal, run:

./gradlew shadowJar

Now that an uberjar for the Kafka Streams application has been built, you can launch it locally. When you run the following, the prompt won’t return, because the application will run until you exit it:

java -jar build/libs/kstreams-merge-standalone-0.0.1.jar configuration/dev.properties

Produce events to the input topics

To produce the input events to their respective topics, you’ll want two terminals running. To send the rock songs to their topic, open up a terminal and run the following:

confluent kafka topic produce rock-song-events \
      --value-format avro \
      --schema src/main/avro/song_event.avsc

You will be prompted for the Confluent Cloud Schema Registry credentials as shown below, which you can find in the configuration/ccloud.properties configuration file. Look for the configuration parameter basic.auth.user.info, whereby the ":" is the delimiter between the key and secret.

Enter your Schema Registry API key:
Enter your Schema Registry API secret:

When the console producer starts, it will log some messages and hang, waiting for your input. Type in one line at a time and press enter to send it. Each line represents an event. To send all of the events below, paste the following into the prompt and press enter:

{"artist": "Metallica", "title": "Fade to Black"}
{"artist": "Smashing Pumpkins", "title": "Today"}
{"artist": "Pink Floyd", "title": "Another Brick in the Wall"}
{"artist": "Van Halen", "title": "Jump"}
{"artist": "Led Zeppelin", "title": "Kashmir"}

To produce the classical songs, open up another terminal and run:

confluent kafka topic produce classical-song-events \
    --value-format avro \
    --schema src/main/avro/song_event.avsc

Then paste in the following events:

{"artist": "Wolfgang Amadeus Mozart", "title": "The Magic Flute"}
{"artist": "Johann Pachelbel", "title": "Canon"}
{"artist": "Ludwig van Beethoven", "title": "Symphony No. 5"}
{"artist": "Edward Elgar", "title": "Pomp and Circumstance"}

Consume events from the output topic

Leaving your original terminals running, open another to consume the events that have been merged:

confluent kafka topic consume all-song-events \
      --from-beginning \
      --value-format avro

After the consumer starts, you should see the following messages. The order might vary depending on the timing of which the input events are sent to each topic and processed by the app. Kafka Streams will coalesce the respective input topics together in an indeterminate manner. To continue studying the example, send more events through the input terminal prompt. Otherwise, you can Control-C to exit the process.

{"artist":"Metallica","title":"Fade to Black"}
{"artist":"Smashing Pumpkins","title":"Today"}
{"artist":"Pink Floyd","title":"Another Brick in the Wall"}
{"artist":"Van Halen","title":"Jump"}
{"artist":"Led Zeppelin","title":"Kashmir"}
{"artist":"Wolfgang Amadeus Mozart","title":"The Magic Flute"}
{"artist":"Johann Pachelbel","title":"Canon"}
{"artist":"Ludwig van Beethoven","title":"Symphony No. 5"}
{"artist":"Edward Elgar","title":"Pomp and Circumstance"}

Teardown Confluent Cloud resources

You may try another tutorial, but if you don’t plan on doing other tutorials, use the Confluent Cloud Console or CLI to destroy all of the resources you created. Verify they are destroyed to avoid unexpected charges.

Test it

Create a test configuration file

First, create a test file at configuration/test.properties:

application.id=merging-app
bootstrap.servers=127.0.0.1:29092
schema.registry.url=mock://merging-app:8081

input.rock.topic.name=rock-song-events
input.rock.topic.partitions=1
input.rock.topic.replication.factor=1

input.classical.topic.name=classical-song-events
input.classical.topic.partitions=1
input.classical.topic.replication.factor=1

output.topic.name=all-song-events
output.topic.partitions=1
output.topic.replication.factor=1

Write a test

Then, create a directory for the tests to live in:

mkdir -p src/test/java/io/confluent/developer

Create the following test file at src/test/java/io/confluent/developer/MergeStreamsTest.java:

package io.confluent.developer;

import org.apache.kafka.common.serialization.Deserializer;
import org.apache.kafka.common.serialization.Serdes;
import org.apache.kafka.common.serialization.Serializer;
import org.apache.kafka.streams.TestInputTopic;
import org.apache.kafka.streams.Topology;
import org.apache.kafka.streams.TopologyTestDriver;
import org.apache.kafka.streams.StreamsConfig;
import org.junit.After;
import org.junit.Assert;
import org.junit.Test;

import java.io.IOException;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Properties;
import java.util.stream.Collectors;

import io.confluent.developer.avro.SongEvent;
import io.confluent.kafka.streams.serdes.avro.SpecificAvroDeserializer;
import io.confluent.kafka.streams.serdes.avro.SpecificAvroSerializer;

public class MergeStreamsTest {

    private final static String TEST_CONFIG_FILE = "configuration/test.properties";
    private TopologyTestDriver testDriver;

    public SpecificAvroSerializer<SongEvent> makeSerializer(Properties allProps) {
        SpecificAvroSerializer<SongEvent> serializer = new SpecificAvroSerializer<>();

        Map<String, String> config = new HashMap<>();
        config.put("schema.registry.url", allProps.getProperty("schema.registry.url"));
        serializer.configure(config, false);

        return serializer;
    }

    public SpecificAvroDeserializer<SongEvent> makeDeserializer(Properties allProps) {
        SpecificAvroDeserializer<SongEvent> deserializer = new SpecificAvroDeserializer<>();

        Map<String, String> config = new HashMap<>();
        config.put("schema.registry.url", allProps.getProperty("schema.registry.url"));
        deserializer.configure(config, false);

        return deserializer;
    }

    @Test
    public void testMergeStreams() throws IOException {
        MergeStreams ms = new MergeStreams();
        Properties allProps = ms.loadEnvProperties(TEST_CONFIG_FILE);
        allProps.put(StreamsConfig.DEFAULT_KEY_SERDE_CLASS_CONFIG, Serdes.String().getClass().getName());
        allProps.put(StreamsConfig.DEFAULT_VALUE_SERDE_CLASS_CONFIG, Serdes.String().getClass().getName());

        String rockTopic = allProps.getProperty("input.rock.topic.name");
        String classicalTopic = allProps.getProperty("input.classical.topic.name");
        String allGenresTopic = allProps.getProperty("output.topic.name");

        Topology topology = ms.buildTopology(allProps);
        testDriver = new TopologyTestDriver(topology, allProps);

        Serializer<String> keySerializer = Serdes.String().serializer();
        SpecificAvroSerializer<SongEvent> valueSerializer = makeSerializer(allProps);

        Deserializer<String> keyDeserializer = Serdes.String().deserializer();
        SpecificAvroDeserializer<SongEvent> valueDeserializer = makeDeserializer(allProps);

        List<SongEvent> rockSongs = new ArrayList<>();
        List<SongEvent> classicalSongs = new ArrayList<>();

        rockSongs.add(SongEvent.newBuilder().setArtist("Metallica").setTitle("Fade to Black").build());
        rockSongs.add(SongEvent.newBuilder().setArtist("Smashing Pumpkins").setTitle("Today").build());
        rockSongs.add(SongEvent.newBuilder().setArtist("Pink Floyd").setTitle("Another Brick in the Wall").build());
        rockSongs.add(SongEvent.newBuilder().setArtist("Van Halen").setTitle("Jump").build());
        rockSongs.add(SongEvent.newBuilder().setArtist("Led Zeppelin").setTitle("Kashmir").build());

        classicalSongs.add(SongEvent.newBuilder().setArtist("Wolfgang Amadeus Mozart").setTitle("The Magic Flute").build());
        classicalSongs.add(SongEvent.newBuilder().setArtist("Johann Pachelbel").setTitle("Canon").build());
        classicalSongs.add(SongEvent.newBuilder().setArtist("Ludwig van Beethoven").setTitle("Symphony No. 5").build());
        classicalSongs.add(SongEvent.newBuilder().setArtist("Edward Elgar").setTitle("Pomp and Circumstance").build());

        final TestInputTopic<String, SongEvent>
            rockSongsTestDriverTopic =
            testDriver.createInputTopic(rockTopic, keySerializer, valueSerializer);

        final TestInputTopic<String, SongEvent>
            classicRockSongsTestDriverTopic =
            testDriver.createInputTopic(classicalTopic, keySerializer, valueSerializer);

        for (SongEvent song : rockSongs) {
            rockSongsTestDriverTopic.pipeInput(song.getArtist(), song);
        }

        for (SongEvent song : classicalSongs) {
            classicRockSongsTestDriverTopic.pipeInput(song.getArtist(), song);
        }

        List<SongEvent> actualOutput =
            testDriver
                .createOutputTopic(allGenresTopic, keyDeserializer, valueDeserializer)
                .readKeyValuesToList()
                .stream()
                .filter(record -> record.value != null)
                .map(record -> record.value)
                .collect(Collectors.toList());

        List<SongEvent> expectedOutput = new ArrayList<>();
        expectedOutput.addAll(rockSongs);
        expectedOutput.addAll(classicalSongs);

        Assert.assertEquals(expectedOutput, actualOutput);
    }

    @After
    public void cleanup() {
        testDriver.close();
    }
}

Invoke the tests

Now run the test, which is as simple as:

./gradlew test