Flink 1.14.* 版本kafkaSink源码

通过Flink 1.14.*版本kafkaSource源码，了解了kafkaSource的源码，这里一起看一下kafkaSink的源码，

一、如何把sink加入到Flink中

首先先找一个类继承FlinkKafkaProducerBase这个抽象类，之后加入到dataStream中就可以了
如下：

DataStream<Tuple2<String, Integer>> dataStream = source的数据
dataStream..addSink(继承了FlinkKafkaProducerBase的实现类)

二、FlinkKafkaProducerBase的源码

@Internal
public abstract class FlinkKafkaProducerBase<IN> extends RichSinkFunction<IN> implements CheckpointedFunction {
    public FlinkKafkaProducerBase(String defaultTopicId, KeyedSerializationSchema<IN> serializationSchema, Properties producerConfig, FlinkKafkaPartitioner<IN> customPartitioner) {
        //省略
    }
    //这个是关键，这个是发送的执行逻辑
    public void invoke(IN next, Context context) throws Exception {
        byte[] serializedKey = this.schema.serializeKey(next);
        byte[] serializedValue = this.schema.serializeValue(next);
        String targetTopic = this.schema.getTargetTopic(next);

        if (this.flinkKafkaPartitioner == null) {
            record = new ProducerRecord(targetTopic, serializedKey, serializedValue);
        } else {
            record = new ProducerRecord(targetTopic, this.flinkKafkaPartitioner.partition(next, serializedKey, serializedValue, targetTopic, partitions), serializedKey, serializedValue);
        }

        this.producer.send(record, this.callback);
    }    
}    

其中invoke是父类RichSinkFunction实现的接口SinkFunction中的方法

三、调用SinkFunction的invoke的地方

@Public
public class DataStream<T> {
    public DataStreamSink<T> addSink(SinkFunction<T> sinkFunction) {
        this.transformation.getOutputType();
        if (sinkFunction instanceof InputTypeConfigurable) {
            ((InputTypeConfigurable)sinkFunction).setInputType(this.getType(), this.getExecutionConfig());
        }
        //这个生成sink算子
        StreamSink<T> sinkOperator = new StreamSink((SinkFunction)this.clean(sinkFunction));
        DataStreamSink<T> sink = new DataStreamSink(this, sinkOperator);
        this.getExecutionEnvironment().addOperator(sink.getLegacyTransformation());
        return sink;
    }

} 

和kafkaSource一样，会把sinkFunction赋值给AbstractUdfStreamOperator的userFunction字段

public class StreamSink<IN> extends AbstractUdfStreamOperator<Object, SinkFunction<IN>> implements OneInputStreamOperator<IN, Object> {
    public StreamSink(SinkFunction<IN> sinkFunction) {
        super(sinkFunction);
        this.chainingStrategy = ChainingStrategy.ALWAYS;
    }
    public void processElement(StreamRecord<IN> element) throws Exception {
        this.sinkContext.element = element;
        ((SinkFunction)this.userFunction).invoke(element.getValue(), this.sinkContext);
    }
}

@PublicEvolving
public abstract class AbstractUdfStreamOperator<OUT, F extends Function> extends AbstractStreamOperator<OUT> implements OutputTypeConfigurable<OUT> {
    private static final long serialVersionUID = 1L;
    protected final F userFunction;

    public AbstractUdfStreamOperator(F userFunction) {
        this.userFunction = (Function)Objects.requireNonNull(userFunction);
        this.checkUdfCheckpointingPreconditions();
    }
}

这里StreamSink也是实现了implements OneInputStreamOperator接口，等Flink运行时调用StreamSink的processElement方法触发实现类的invoke方法，

下面是FlinkKafkaProducerBase类关系图