1、9999端口未占用
netstat -apn|grep 9999
2、消息发送端
nc -l -k -p 9999{"user":"ming","url":"www.baidu1.com","timestamp":1200L}
{"user":"xiaohu","url":"www.baidu5.com","timestamp":1267L}
{"user":"ming","url":"www.baidu7.com","timestamp":4200L}
{"user":"xiaohu","url":"www.baidu8.com","timestamp":5500L}
{"user":"Biu","url":"www.baidu8.com","timestamp":5500L}
3、运行
周期性水位线
import com.alibaba.fastjson2.JSONObject;
import org.apache.flink.api.common.eventtime.*;
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.common.typeinfo.BasicTypeInfo;
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 java.sql.Timestamp;
import java.util.ArrayList;/*** Description: * forMonotonousTimestamps->AscendingTimestampsWatermarks 有序流 -> 自定义断点式水位线（周期延迟时间=0ms）\* forBoundedOutOfOrderness->BoundedOutOfOrdernessWatermarks 无序流 -> 自定义周期性水位线*/
public class FlinkPeriodicWatermarkGeneratorTestJob {public static void main(String[] args) throws Exception {StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();// ArrayList<Event> list = new ArrayList<>();
// list.add(new Event("ming","www.baidu1.com",1200L));
// list.add(new Event("xiaohu","www.baidu5.com",1267L));
// list.add(new Event("ming","www.baidu7.com",4200L));
// list.add(new Event("xiaohu","www.baidu8.com",5500L));
//
// DataStreamSource<Event> ds = env.fromCollection(list, BasicTypeInfo.of(Event.class));DataStreamSource<String> dss = env.socketTextStream("test002", 9999);SingleOutputStreamOperator<Event> ds = dss.map(new MapFunction<String, Event>() {@Overridepublic Event map(String value) throws Exception {Event event = new Event();event.toEvent(value);return event;}});
// ds.print();SingleOutputStreamOperator<Event> watermarks = ds// AscendingTimestampsWatermarks 有序流 查看源码，实际上是延迟时间=0ms的乱序流
// .assignTimestampsAndWatermarks(WatermarkStrategy.<Event>forMonotonousTimestamps()// BoundedOutOfOrdernessWatermarks 无序流 5ms固定延迟时间/表示最大乱序程度 处理乱序流数据.assignTimestampsAndWatermarks(new WatermarkStrategy<Event>() {@Overridepublic TimestampAssigner<Event> createTimestampAssigner(TimestampAssignerSupplier.Context context) {return new SerializableTimestampAssigner<Event>() {@Overridepublic long extractTimestamp(Event element, long recordTimestamp) {return element.getTimestamp();}};}@Overridepublic WatermarkGenerator<Event> createWatermarkGenerator(WatermarkGeneratorSupplier.Context context) {return new WatermarkGenerator<Event>() {private Long delayTime = 5000L; // 延迟时间private Long maxTs = Long.MIN_VALUE + delayTime + 1L;@Overridepublic void onEvent(Event event, long eventTimestamp, WatermarkOutput output) {// 每来一条数据就调用一次maxTs = Math.max(event.timestamp, maxTs);// 更新最大时间戳}@Overridepublic void onPeriodicEmit(WatermarkOutput output) {// 发射水位线，默认 200ms 调用一次 可以使用 env.getConfig().setAutoWatermarkInterval(60 * 1000L); 调整周期时间 flink时间窗口(左开,右闭]output.emitWatermark(new Watermark(maxTs - delayTime - 1L));}};}});ds.print();env.setParallelism(1);env.execute();}public static class Event{String user;String url;Long timestamp;public Event(){}public Event(String user, String url, Long timestamp) {this.user = user;this.url = url;this.timestamp = timestamp;}public String getUser() {return user;}public String getUrl() {return url;}public Long getTimestamp() {return timestamp;}@Overridepublic String toString() {return "Event{" +"user='" + user + '\'' +", url='" + url + '\'' +", timestamp=" + new Timestamp(timestamp) +'}';}public void toEvent(String val){JSONObject js = JSONObject.parseObject(val);this.user = js.getString("user");this.url = js.getString("url");this.timestamp = js.getLong("timestamp");}}
}断点式水位线
import com.alibaba.fastjson2.JSONObject;
import org.apache.flink.api.common.eventtime.*;
import org.apache.flink.api.common.functions.MapFunction;
import org.apache.flink.api.common.typeinfo.BasicTypeInfo;
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 java.sql.Timestamp;
import java.util.ArrayList;/*** Description: * forMonotonousTimestamps->AscendingTimestampsWatermarks 有序流 -> 自定义断点式水位线（周期延迟时间=0ms）\* forBoundedOutOfOrderness->BoundedOutOfOrdernessWatermarks 无序流 -> 自定义周期性水位线*/
public class FlinkPunctuatedWatermarkGeneratorTestJob {public static void main(String[] args) throws Exception {StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();DataStreamSource<String> dss = env.socketTextStream("test002", 9999);SingleOutputStreamOperator<Event> ds = dss.map(new MapFunction<String, Event>() {@Overridepublic Event map(String value) throws Exception {Event event = new Event();event.toEvent(value);return event;}});
// ds.print();SingleOutputStreamOperator<Event> watermarks = ds// AscendingTimestampsWatermarks 有序流 查看源码，实际上是延迟时间=0ms的乱序流
// .assignTimestampsAndWatermarks(WatermarkStrategy.<Event>forMonotonousTimestamps()// BoundedOutOfOrdernessWatermarks 无序流 5ms固定延迟时间/表示最大乱序程度 处理乱序流数据.assignTimestampsAndWatermarks(new WatermarkStrategy<Event>() {@Overridepublic TimestampAssigner<Event> createTimestampAssigner(TimestampAssignerSupplier.Context context) {return new SerializableTimestampAssigner<Event>() {@Overridepublic long extractTimestamp(Event element, long recordTimestamp) {return element.getTimestamp();}};}@Overridepublic WatermarkGenerator<Event> createWatermarkGenerator(WatermarkGeneratorSupplier.Context context) {return new WatermarkGenerator<Event>() {@Overridepublic void onEvent(Event event, long eventTimestamp, WatermarkOutput output) {// 只有在遇到特定的 itemId 时，才发出水位线if (event.getUser().equals("Biu")) {output.emitWatermark(new Watermark(event.getTimestamp() - 1));}}@Overridepublic void onPeriodicEmit(WatermarkOutput output) {// 不需要做任何事情，因为我们在 onEvent 方法中发射了水位线}};}});ds.print();env.setParallelism(1);env.execute();}public static class Event{String user;String url;Long timestamp;public Event(){}public Event(String user, String url, Long timestamp) {this.user = user;this.url = url;this.timestamp = timestamp;}public String getUser() {return user;}public String getUrl() {return url;}public Long getTimestamp() {return timestamp;}@Overridepublic String toString() {return "Event{" +"user='" + user + '\'' +", url='" + url + '\'' +", timestamp=" + new Timestamp(timestamp) +'}';}public void toEvent(String val){JSONObject js = JSONObject.parseObject(val);this.user = js.getString("user");this.url = js.getString("url");this.timestamp = js.getLong("timestamp");}}
}4、打印
3> Event{user='ming', url='www.baidu1.com', timestamp=1970-01-01 08:00:01.2}
4> Event{user='xiaohu', url='www.baidu5.com', timestamp=1970-01-01 08:00:01.267}
5> Event{user='ming', url='www.baidu7.com', timestamp=1970-01-01 08:00:04.2}
6> Event{user='xiaohu', url='www.baidu8.com', timestamp=1970-01-01 08:00:05.5}参考：
【Flink】Flink 中的时间和窗口之水位线(Watermark)-CSDN博客
Flink watermark_nc -lp 9999-CSDN博客
NoteWarehouse/05_BigData/09_Flink(1).md at main · FGL12321/NoteWarehouse · GitHub