flink 自定义周期水印，解决最后一个窗口因为缺少数据到达，不触发问题

1.自定义周期水印

public class MyWaterMark implements AssignerWithPeriodicWatermarks<MyEvent> {

    private final long maxTimeLag = 3000; // 3 seconds

    @Override
    public long extractTimestamp(MyEvent element, long previousElementTimestamp) {
        return element.getCreationTime();
    }

    @Override
    public Watermark getCurrentWatermark() {
        // return the watermark as current time minus the maximum time lag
        return new Watermark(System.currentTimeMillis() - maxTimeLag);
    }
}

2.使用自定义水印

val env = StreamExecutionEnvironment.getExecutionEnvironment

//便于测试，并行度设置为1
env.setParallelism(1)

//env.getConfig.setAutoWatermarkInterval(9000)

//设置为事件时间
env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)

//设置source 本地socket
val text: DataStream[String] = env.socketTextStream("localhost", 9000)


val lateText = new OutputTag[(String, String, Long, Long)]("late_data")

val value = text.filter(new MyFilterNullOrWhitespace)
.flatMap(new MyFlatMap)
//使用自定义水印
.assignTimestampsAndWatermarks(new MyWaterMark())
.map(x => (x.name, x.datetime, x.timestamp, 1L))
.keyBy(_._1)
.window(TumblingEventTimeWindows.of(Time.seconds(5)))
.sideOutputLateData(lateText)
//.sum(2)
.apply(new MyWindow)
//.window(TumblingEventTimeWindows.of(Time.seconds(3)))
//.apply(new MyWindow)
value.getSideOutput(lateText).map(x => {
"延迟数据|name:" + x._1 + "|datetime:" + x._2
}).print()

value.print()

env.execute("watermark test")

参考：flink Periodic Watermarks 自定义周期性水印 - 我是属车的 - 博客园