CODEWITHSUNDEEP
javaapache-sparkapache-kafkaavro
Claudio Melis
Claudio Melis

Reputation: 23

Java Kafka consumer and avro deserialzier

I'm developing a simple java with spark streaming.

I configured a kafka jdbc connector (postgres to topic) and I wanna read it with a spark streaming consumer.

I'm able to read to topic correctly with:

./kafka-avro-console-consumer --bootstrap-server localhost:9092 --property schema.registry.url=http://localhost:8081 --property print.key=true --from-beginning --topic postgres-ip_audit

getting this results:

null {"id":1557,"ip":{"string":"90.228.176.138"},"create_ts":{"long":1554819937582}}

when I use my java application with this config:

Map<String, Object> kafkaParams = new HashMap<>();
kafkaParams.put("bootstrap.servers", "localhost:9092");
kafkaParams.put("key.deserializer", StringDeserializer.class);
kafkaParams.put("value.deserializer", StringDeserializer.class);
kafkaParams.put("group.id", "groupStreamId");
kafkaParams.put("auto.offset.reset", "latest");
kafkaParams.put("enable.auto.commit", false);

I get results like that:

�179.20.119.53�����Z

Can someone point me how to fix my issue?

I try also to use a ByteArrayDeserializer and convert the bytes[] in to a string but I get always bad character results.

Upvotes: 2

Views: 6710

Answers (2)

Somasundaram Sekar
Somasundaram Sekar

Reputation: 5524

You can deserialize avro messages using io.confluent.kafka.serializers.KafkaAvroDeserializer and having schema registry in to manage the records schema.

Here is a sample code snippet 

import java.util.Collections;
import java.util.HashMap;
import java.util.HashSet;
import java.util.Map;
import java.util.Set;

import io.confluent.kafka.serializers.KafkaAvroDecoder;
import kafka.serializer.StringDecoder;
import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.streaming.Durations;
import org.apache.spark.streaming.api.java.JavaPairInputDStream;
import org.apache.spark.streaming.api.java.JavaStreamingContext;
import org.apache.spark.streaming.kafka.KafkaUtils;
import scala.Tuple2;

public class SparkStreaming {

  public static void main(String... args) {
    SparkConf conf = new SparkConf();
    conf.setMaster("local[2]");
    conf.setAppName("Spark Streaming Test Java");

    JavaSparkContext sc = new JavaSparkContext(conf);
    JavaStreamingContext ssc = new JavaStreamingContext(sc, Durations.seconds(10));

    processStream(ssc, sc);

    ssc.start();
    ssc.awaitTermination();
  }

  private static void processStream(JavaStreamingContext ssc, JavaSparkContext sc) {
    System.out.println("--> Processing stream");

    Map<String, String> props = new HashMap<>();
    props.put("bootstrap.servers", "localhost:9092");
    props.put("schema.registry.url", "http://localhost:8081");
    props.put("group.id", "spark");
    props.put("specific.avro.reader", "true");

    props.put("value.deserializer", "io.confluent.kafka.serializers.KafkaAvroDeserializer");
    props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");

    Set<String> topicsSet = new HashSet<>(Collections.singletonList("test"));

    JavaPairInputDStream<String, Object> stream = KafkaUtils.createDirectStream(ssc, String.class, Object.class,
      StringDecoder.class, KafkaAvroDecoder.class, props, topicsSet);

    stream.foreachRDD(rdd -> {
      rdd.foreachPartition(iterator -> {
          while (iterator.hasNext()) {
            Tuple2<String, Object> next = iterator.next();
            Model model = (Model) next._2();
            System.out.println(next._1() + " --> " + model);
          }
        }
      );
    });
  }
}

Complete sample application is available in this github repo

Upvotes: 1

JEY
JEY

Reputation: 7133

You provided a StringDeserializer however you are sending values serialized with avro so you need to deserialized them accordingly. Using spark 2.4.0 (and the following deps compile org.apache.spark:spark-avro_2.12:2.4.1 you can achieve it by using from_avro function:

import org.apache.spark.sql.avro._

// `from_avro` requires Avro schema in JSON string format.
val jsonFormatSchema = new String(Files.readAllBytes(Paths.get("path/to/your/schema.avsc")))

val df = spark
  .readStream
  .format("kafka")
  .option("kafka.bootstrap.servers", "host1:port1,host2:port2")
  .option("subscribe", "topic1")
  .load()
Dataset<Row> output = df
  .select(from_avro(col("value"), jsonFormatSchema).as("user"))
  .where("user.favorite_color == \"red\"")
  .show()

If you need to use a schema registry (like you did with kafka-avro-console-consumer) it's not possible out of the box and need a to write a lot of code. I'll recommend using this lib https://github.com/AbsaOSS/ABRiS. However it's only compatible with spark 2.3.0

Upvotes: 0

Related Questions