你知道flink侧输出流产生分流，和直接DataStream使用多次产生分流的区别吗？

应该说有两个区别：

1：侧输出流的输出类型可以是任意的，但是一条流处理多次的时候一定会经历filter阶段，在经历fliter阶段输出类型一定是相同。

2：一条流处理多次的时候相当于使用了多个if操作。但是侧输出流，可以使用if、else if、else这样的操作，如果if、else if满足条件，可以提前结束方法，减少某些数据的逻辑执行。

结果上看是没有区别的，但是从实现上来说，一般旁路输出只需要对一条流处理一次(类似于split)，但是两条流一般都是对流filter，这样就相当于对一条流处理了两次才能生成两条流。

看一个代码例子，大家瞬间就明白了：

首先来一个一个流执行多次的代码案例（一定有多次filter）

package com.andy;

import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.java.tuple.Tuple4;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;


public class SideOutputTest {
    /**
     * 整个方法假设只在分别统计访问量最多的10个用户，以及访问量最多的10个url，以及访问量最多的10个ip
     *
     * @param args
     */
    public static void main(String[] args) {
        StreamExecutionEnvironment environment = StreamExecutionEnvironment.getExecutionEnvironment();
        //假设每行的数据是businessType,userId,url,ip(中间以逗号隔开,假设businessType值又jd,taobao)
        DataStreamSource<String> dataStreamSource = environment.socketTextStream("localhost", 9999);

        SingleOutputStreamOperator<Tuple4<String, String, String, String>> afterMap =
                dataStreamSource.map(new MapFunction<String, Tuple4<String, String, String, String>>() {

                    @Override
                    public Tuple4<String, String, String, String> map(String input) throws Exception {
                        String[] userIdUrlIp = input.split(",");
                        return Tuple4.of(userIdUrlIp[0], userIdUrlIp[1], userIdUrlIp[2], userIdUrlIp[3]);
                    }
                });

        afterMap.filter(item -> item.f0.equalsIgnoreCase("jd"))
                .print("jd");
        afterMap.filter(item -> item.f0.equalsIgnoreCase("taobao"))
                .print("taobao");
        try {
            environment.execute();
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}

再来一个侧输出流的例子：

package com.andy;

import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.java.tuple.Tuple3;
import org.apache.flink.api.java.tuple.Tuple4;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.ProcessFunction;
import org.apache.flink.util.Collector;
import org.apache.flink.util.OutputTag;


public class SideOutputTest2 {
    /**
     * 整个方法假设只在分别统计访问量最多的10个用户，以及访问量最多的10个url，以及访问量最多的10个ip
     *
     * @param args
     */
    public static void main(String[] args) {
        StreamExecutionEnvironment environment = StreamExecutionEnvironment.getExecutionEnvironment();
        //假设每行的数据是businessType,userId,url,ip(中间以逗号隔开,假设businessType值又jd,taobao)
        DataStreamSource<String> dataStreamSource = environment.socketTextStream("localhost", 9999);

        SingleOutputStreamOperator<Tuple4<String, String, String, String>> afterMap =
                dataStreamSource.map(new MapFunction<String, Tuple4<String, String, String, String>>() {

                    @Override
                    public Tuple4<String, String, String, String> map(String input) throws Exception {
                        String[] userIdUrlIp = input.split(",");
                        return Tuple4.of(userIdUrlIp[0], userIdUrlIp[1], userIdUrlIp[2], userIdUrlIp[3]);
                    }
                });

        //侧输出流的时候，类型在processElement里面就可以从Tuple4变成Tuple3
        OutputTag<Tuple3<String, String, String>> jd = new OutputTag<Tuple3<String, String, String>>("jd");
        OutputTag<Tuple3<String, String, String>> taobao = new OutputTag<Tuple3<String, String, String>>("taobao");
        SingleOutputStreamOperator<Object> processedDataStream = afterMap.process(new ProcessFunction<Tuple4<String, String, String, String>, Object>() {
            @Override
            public void processElement(Tuple4<String, String, String, String> input, ProcessFunction<Tuple4<String, String, String, String>, Object>.Context ctx, Collector<Object> out) throws Exception {
                if (input.f0.equalsIgnoreCase("jd")) {
                    ctx.output(jd, Tuple3.of(input.f1, input.f2, input.f3));
                } else if (input.f0.equalsIgnoreCase("taobao")) {
                    ctx.output(taobao, Tuple3.of(input.f1, input.f2, input.f3));
                } else {
                    out.collect(input);
                }
            }
        });
        processedDataStream.getSideOutput(jd).print("jd");
        processedDataStream.getSideOutput(taobao).print("taobao");
        processedDataStream.print("except jd and taobao");
        try {
            environment.execute();
        } catch (Exception e) {
            e.printStackTrace();
        }
    }
}