1.首先自定义个 KafkaDeserializationSchema

public class CustomKafkaDeserializationSchema implements KafkaDeserializationSchema<Tuple2<String, String>> {
    @Override
    //nextElement 是否表示流的最后一条元素，我们要设置为 false ,因为我们需要 msg 源源不断的被消费
    public boolean isEndOfStream(Tuple2<String, String> nextElement) {
        return false;
    }
    
    @Override
    // 反序列化 kafka 的 record，我们直接返回一个 tuple2<kafkaTopicName,kafkaMsgValue>
    public Tuple2<String, String> deserialize(ConsumerRecord<byte[], byte[]> record) throws Exception {
        return new Tuple2<>(record.topic(), new String(record.value(), "UTF-8"));
    }
    
    @Override
    //告诉 Flink 我输入的数据类型, 方便 Flink 的类型推断
    public TypeInformation<Tuple2<String, String>> getProducedType() {
        return new TupleTypeInfo<>(BasicTypeInfo.STRING_TYPE_INFO, BasicTypeInfo.STRING_TYPE_INFO);
    }
}

2.使用自定义的 KafkaDeserializationSchema 进行消费

public static void main(String[] args) throws Exception {
        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
                
        Properties properties = new Properties();
        properties.setProperty("bootstrap.servers", "localhost:9092");
        properties.setProperty("group.id", "test");
        
        FlinkKafkaConsumer<Tuple2<String, String>> kafkaConsumer = new FlinkKafkaConsumer<>("test", new CustomKafkaDeserializationSchema(), properties);
        kafkaConsumer.setStartFromEarliest();
        env.addSource(kafkaConsumer).flatMap(new FlatMapFunction<Tuple2<String, String>, Object>() {
            @Override
            public void flatMap(Tuple2<String, String> value, Collector<Object> out) throws Exception {
                System.out.println("topic==== " + value.f0);
            }
        });
        
        // execute program
        env.execute("Flink Streaming Java API Skeleton");
    }