Deserialise a POJO in Kafka Streams

问题:

My Kafka topic has messages of this format

user1,subject1,80|user1,subject2,90 

user2,subject1,70|user2,subject2,100 

and so on. 

I have created User POJO as below.

class User implements Serializable{
/**
 * 
 */
private static final long serialVersionUID = -253687203767610477L;
private String userId;
private String subject;
private String marks;

public User(String userId, String subject, String marks) {
    super();
    this.userId = userId;
    this.subject = subject;
    this.marks = marks;
}

public String getUserId() {
    return userId;
}

public void setUserId(String userId) {
    this.userId = userId;
}
public String getSubject() {
    return subject;
}
public void setSubject(String subject) {
    this.subject = subject;
}
public String getMarks() {
    return marks;
}
public void setMarks(String marks) {
    this.marks = marks;
}
}

Further I have created default key value serialization

streamProperties.put(
            StreamsConfig.DEFAULT_KEY_SERDE_CLASS_CONFIG, Serdes.String().getClass());
streamProperties.put(
            StreamsConfig.DEFAULT_VALUE_SERDE_CLASS_CONFIG, Serdes.String().getClass());

I am trying to find count by userID as follows. Also I need User object to perform some other functionalities.

KTable<String, Long> wordCount = streamInput

    .flatMap(new KeyValueMapper<String, String, Iterable<KeyValue<String,User>>>() {

        @Override
        public Iterable<KeyValue<String, User>> apply(String key, String value) {
            String[] userObjects = value.split("|");
            List<KeyValue<String, User>> userList = new LinkedList<>();
            for(String userObject: userObjects) {
                String[] userData = userObject.split(",");
                userList.add(KeyValue.pair(userData[0],
                        new User(userData[0],userData[1],userData[2])));


            }
            return userList;
        }
    })

.groupByKey()
.count();

I am getting the below error

Caused by: org.apache.kafka.streams.errors.StreamsException: A serializer (key: org.apache.kafka.common.serialization.StringSerializer / value: org.apache.kafka.common.serialization.StringSerializer) is not compatible to the actual key or value type (key type: java.lang.String / value type: com.example.testing.dao.User). Change the default Serdes in StreamConfig or provide correct Serdes via method parameters.

I think I need to provide correct Serde for User Class.

回答1:

The problem is with Value Serdes.

There are two version of function groupBy:

• KStream::KGroupedStream<K, V> groupByKey();
• KStream::KGroupedStream<K, V> groupByKey(final Grouped<K, V> grouped);

First version under the hood call second with Grouped with default Serdes (In your case it was for key and value StringSerde

Your flatMap map message to KeyValue<String, User> type so value was of type User.

Solution in your case would be instead using groupByKey() call groupByKey(Grouped.with(keySerde, valSerde));, with proper Serdes.