Flink实战：自定义KafkaDeserializationSchema(Java/Scala)

微信公众号：大数据开发运维架构

关注可了解更多大数据相关的资讯。问题或建议，请公众号留言;

如果您觉得“大数据开发运维架构”对你有帮助，欢迎转发朋友圈

从微信公众号拷贝过来，格式有些错乱，建议直接去公众号阅读

---

    kafka中的数据通常是键值对的，所以我们这里自定义反序列化类从kafka中消费键值对的消息，为方便大家学习，这里我实现了Java/Scala两个版本，由于比较简单这里直接上代码：

一、Scala代码：

1.自定义反序列化类：

package comhadoop.ljs.flink010.kafka
import org.apache.flink.api.common.typeinfo.{TypeHint, TypeInformation}
import org.apache.flink.streaming.connectors.kafka.KafkaDeserializationSchema
import org.apache.kafka.clients.consumer.ConsumerRecord
/**
  * @author: Created By lujisen
  * @company ChinaUnicom Software JiNan
  * @date: 2020-04-25 18:31
  * @version: v1.0
  * @description: comhadoop.ljs.flink010.kafka
  */
class MyKafkaDeserializationSchema  extends KafkaDeserializationSchema[ConsumerRecord[String, String]]{
  /*是否流结束，比如读到一个key为end的字符串结束，这里不再判断，直接返回false 不结束*/
  override def isEndOfStream(t: ConsumerRecord[String, String]): Boolean ={
    false
  }
  override def deserialize(record: ConsumerRecord[Array[Byte], Array[Byte]]): ConsumerRecord[String, String] = {
    new ConsumerRecord(record.topic(),record.partition(),record.offset(),new String(record.key(),"UTF-8"),new String(record.value(),"UTF-8"))
  }
  /*用于获取反序列化对象的类型*/
  override def getProducedType: TypeInformation[ConsumerRecord[String, String]] = {
    TypeInformation.of(new TypeHint[ConsumerRecord[String, String]] {})
  }
}

2.主函数类：

package comhadoop.ljs.flink010.kafka
import java.util.Properties
import org.apache.flink.api.common.functions.MapFunction
import org.apache.flink.streaming.api.datastream.DataStream
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer
import org.apache.flink.streaming.connectors.kafka.internals.KafkaTopicPartition
import org.apache.kafka.clients.consumer.ConsumerRecord
import org.apache.kafka.common.serialization.StringDeserializer
/**
  * @author: Created By lujisen
  * @company ChinaUnicom Software JiNan
  * @date: 2020-04-25 16:32
  * @version: v1.0
  * @description: comhadoop.ljs.flink010.kafka
  */
object KafkaDeserializerSchemaTest {
  def main(args: Array[String]): Unit = {

    /*环境初始化*/
    val senv:StreamExecutionEnvironment =StreamExecutionEnvironment.getExecutionEnvironment()
    /*启用checkpoint，这里我没有对消息体的key value进行判断，即使为空启动了checkpoint，遇到错误也会无限次重启*/
    senv.enableCheckpointing(2000)
    /*topic2不存在话会自动在kafka创建，一个分区 分区名称0*/
    val  myConsumer=new FlinkKafkaConsumer[ConsumerRecord[String, String]]("topic3",new MyKafkaDeserializationSchema(),getKafkaConfig())
    /*指定消费位点*/
    val specificStartOffsets = new java.util.HashMap[KafkaTopicPartition, java.lang.Long]()
    /*这里从topic3 的0分区的第一条开始消费*/
    specificStartOffsets.put(new KafkaTopicPartition("topic3", 0), 0L)
    myConsumer.setStartFromSpecificOffsets(specificStartOffsets)
    /*指定source数据源*/
    val source:DataStream[ConsumerRecord[String, String]]=senv.addSource(myConsumer)

    val keyValue=source.map(new MapFunction[ConsumerRecord[String, String],String] {
      override def map(message: ConsumerRecord[String, String]): String = {
        "key" + message.key + "  value:" + message.value
      }
    })
    /*打印接收的数据*/
    keyValue.print()
    /*启动执行*/
    senv.execute()
  }

  def getKafkaConfig():Properties={
    val props:Properties=new Properties()
    props.setProperty("bootstrap.servers","worker1.hadoop.ljs:6667,worker2.hadoop.ljs:6667")
    props.setProperty("group.id","topic_1")
    props.setProperty("key.deserializer",classOf[StringDeserializer].getName)
    props.setProperty("value.deserializer",classOf[StringDeserializer].getName)
    props.setProperty("auto.offset.reset","latest")
    props
  }
}

二、Java代码： 
1.自定义反序列化类：

package com.hadoop.ljs.flink110.kafka;
import org.apache.flink.api.common.typeinfo.TypeHint;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.streaming.connectors.kafka.KafkaDeserializationSchema;
import org.apache.kafka.clients.consumer.ConsumerRecord;
/**
 * @author: Created By lujisen
 * @company ChinaUnicom Software JiNan
 * @date: 2020-04-25 18:45
 * @version: v1.0
 * @description: com.hadoop.ljs.flink110.kafka
 */
public class MyKafkaDeserializationSchema implements KafkaDeserializationSchema> {

    private static  String encoding = "UTF8";
    @Override
    public boolean isEndOfStream(ConsumerRecord nextElement) {
        return false;
    }
    @Override
    public ConsumerRecord deserialize(ConsumerRecord record) throws Exception {
       /* System.out.println("Record--partition::"+record.partition());
        System.out.println("Record--offset::"+record.offset());
        System.out.println("Record--timestamp::"+record.timestamp());
        System.out.println("Record--timestampType::"+record.timestampType());
        System.out.println("Record--checksum::"+record.checksum());
        System.out.println("Record--key::"+record.key());
        System.out.println("Record--value::"+record.value());*/
        return new ConsumerRecord(record.topic(),
                record.partition(),
                record.offset(),
                record.timestamp(),
                record.timestampType(),
                record.checksum(),
                record.serializedKeySize(),
                record.serializedValueSize(),
                /*这里我没有进行空值判断，生产一定记得处理*/
                new  String(record.key(), encoding),
                new  String(record.value(), encoding));
    }
    @Override
    public TypeInformation> getProducedType() {
        return TypeInformation.of(new TypeHint>(){});
    }
}

 

2.主函数类：

package com.hadoop.ljs.flink110.kafka;
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer;
import org.apache.flink.streaming.connectors.kafka.internals.KafkaTopicPartition;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import java.util.HashMap;
import java.util.Map;
import java.util.Properties;
/**
 * @author: Created By lujisen
 * @company ChinaUnicom Software JiNan
 * @date: 2020-04-25 18:41
 * @version: v1.0
 * @description: com.hadoop.ljs.flink110.kafka
 */
public class KafkaDeserializerSchemaTest {
    public static void main(String[] args) throws Exception {
        /*环境初始化*/
        StreamExecutionEnvironment senv = StreamExecutionEnvironment.getExecutionEnvironment();
        /*启用checkpoint，这里我没有对消息体的key value进行判断，即使为空启动了checkpoint，遇到错误也会无限次重启*/
        senv.enableCheckpointing(2000);
        /*topic2不存在话会自动在kafka创建，一个分区 分区名称0*/
        FlinkKafkaConsumer> myConsumer=new FlinkKafkaConsumer>("topic3",new MyKafkaDeserializationSchema(),getKafkaConfig());

        /*指定消费位点*/
        Map specificStartOffsets = new HashMap<>();
        /*这里从topic3 的0分区的第一条开始消费*/
        specificStartOffsets.put(new KafkaTopicPartition("topic3", 0), 0L);
        myConsumer.setStartFromSpecificOffsets(specificStartOffsets);

        DataStream> source = senv.addSource(myConsumer);
        DataStream keyValue = source.map(new MapFunction, String>() {
            @Override
            public String map(ConsumerRecord message) throws Exception {
                return "key"+message.key()+"  value:"+message.value();
            }
        });
        /*打印结果*/
        keyValue.print();
        /*启动执行*/
        senv.execute();
    }
    public static Properties getKafkaConfig(){
        Properties  props=new Properties();
        props.setProperty("bootstrap.servers","worker1.hadoop.ljs:6667,worker2.hadoop.ljs:6667");
        props.setProperty("group.id","topic_group2");
        props.put("key.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        props.put("value.deserializer", "org.apache.kafka.common.serialization.StringDeserializer");
        props.setProperty("auto.offset.reset","latest");
        return props;
    }
}

三、函数测试

1.KafkaProducer发送测试数据类：

package com.hadoop.ljs.kafka220;
import org.apache.kafka.clients.producer.KafkaProducer;
import org.apache.kafka.clients.producer.ProducerRecord;
import java.util.Date;
import java.util.Properties;
public class KafkaPartitionProducer extends Thread{
  private static long count =10;
  private static String topic="topic3";
  private static String brokerList="worker1.hadoop.ljs:6667,worker2.hadoop.ljs:6667";
  public static void main(String[] args) {
        KafkaPartitionProducer jproducer = new KafkaPartitionProducer();
        jproducer.start();
    }
    @Override
    public void run() {
        producer();
    }
    private void producer() {
        Properties props = config();
        KafkaProducer producer = new KafkaProducer<>(props);
        ProducerRecord record=null;
        System.out.println("kafka生产数据条数："+count);
        for (int i = 1; i <= count; i++) {
            String json = "{\"id\":" + i + ",\"ip\":\"192.168.0." + i + "\",\"date\":" + new Date().toString() + "}";
            String key ="key"+i;
            record = new ProducerRecord(topic, key, json);
            producer.send(record, (metadata, e) -> {
                // 使用回调函数
                if (null != e) {
                    e.printStackTrace();
                }
                if (null != metadata) {
                    System.out.println(String.format("offset: %s, partition:%s, topic:%s  timestamp:%s", metadata.offset(), metadata.partition(), metadata.topic(), metadata.timestamp()));
                }
            });
        }
        producer.close();
    }
    private Properties config() {
        Properties props = new Properties();
        props.put("bootstrap.servers",brokerList);
        props.put("acks", "1");
        props.put("retries", 3);
        props.put("batch.size", 16384);
        props.put("linger.ms", 1);
        props.put("buffer.memory", 33554432);
        props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer");
        props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer");
        /*自定义分区，两种形式*/
        /*props.put("partitioner.class", PartitionUtil.class.getName());*/
        return props;
    }
}

2.测试结果

如果觉得我的文章能帮到您，请关注微信公众号“大数据开发运维架构”，并转发朋友圈，谢谢支持！！！