Flink Window DEMO 學習

該文檔演示了fink windows的操作DEMO

環境準備：

kafka本地運行：kafka部署
自動生成名字代碼：隨機名
自動生成隨機IP代碼：隨機IP
Flink 1.18

測試數據

自動向kafka推送數據

import cn.hutool.core.date.DateUtil;
import com.alibaba.fastjson2.JSONObject;
import com.wfg.flink.example.dto.KafkaPvDto;
import com.wfg.flink.example.utils.RandomGeneratorUtils;
import org.apache.kafka.clients.producer.KafkaProducer;
import org.apache.kafka.clients.producer.Producer;
import org.apache.kafka.clients.producer.ProducerRecord;import java.time.LocalDateTime;
import java.util.Properties;
import java.util.Random;
import java.util.UUID;
import java.util.concurrent.CompletableFuture;import static com.wfg.flink.example.constants.Constants.KAFKA_BROKERS;
import static com.wfg.flink.example.constants.Constants.TOPIC_NAME;public class KafkaTestProducer {public static void main(String[] args) {Properties props = new Properties();props.put("bootstrap.servers", KAFKA_BROKERS);props.put("key.serializer", "org.apache.kafka.common.serialization.StringSerializer");props.put("value.serializer", "org.apache.kafka.common.serialization.StringSerializer");try (Producer<String, String> producer = new KafkaProducer<>(props)) {int times = 100000;for (int i = 0; i < times; i++) {System.out.println("Send No. :" + i);CompletableFuture.allOf(CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer)),CompletableFuture.runAsync(() -> sendKafkaMsg(producer))).join();producer.flush();Random random = new Random();int randomNumber = random.nextInt(7); // 生成一個0到6的隨機數Thread.sleep(1000 * randomNumber);}} catch (InterruptedException e) {throw new RuntimeException(e);}}private static void sendKafkaMsg(Producer<String, String> producer) {String msg = createMsg();System.out.println(msg);producer.send(new ProducerRecord<>(TOPIC_NAME, UUID.randomUUID().toString().replaceAll("-", ""), msg));}private static String createMsg() {KafkaPvDto dto = new KafkaPvDto();dto.setUuid(UUID.randomUUID().toString().replaceAll("-", ""));dto.setUserName(RandomGeneratorUtils.generateRandomFullName());dto.setVisitIp(RandomGeneratorUtils.generateRandomIp());
//        DateTime begin = DateUtil.beginOfDay(new Date());
//        String timeStr = DateUtil.format(RandomGeneratorUtils.generateRandomDateTime(LocalDateTimeUtil.of(begin).toLocalDate(), LocalDate.now()), "yyyy-MM-dd HH:mm:ss");String timeStr = DateUtil.format(LocalDateTime.now(), "yyyy-MM-dd HH:mm:ss");dto.setVisitTime(timeStr);dto.setVisitServiceIp(RandomGeneratorUtils.generateRandomIp());return JSONObject.toJSONString(dto);}
}

注意：

kafka本地運行：kafka部署
自動生成名字代碼：隨機名
自動生成隨機IP代碼：隨機IP

FLINK 數據


/**** @author wfg*/
@Slf4j
public class DataSplitter implements FlatMapFunction<String, Tuple2<String, Integer>> {@Overridepublic void flatMap(String value, Collector<Tuple2<String, Integer>> collector) {KafkaPvDto data = JSONObject.parseObject(value, KafkaPvDto.class);if (data != null) {collector.collect(new Tuple2<>(data.getUserName(), 1));}}
}

基于時間窗口

*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//基于時間窗口data.flatMap(new DataSplitter()).keyBy(1).timeWindow(Time.seconds(30)).sum(0).print();*/env.execute("flink window example");}
}

基于滑動時間窗口

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//基于滑動時間窗口data.flatMap(new DataSplitter()).keyBy(1).timeWindow(Time.seconds(60), Time.seconds(30)).sum(0).print();env.execute("flink window example");}
}

基于事件數量窗口

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//基于事件數量窗口data.flatMap(new DataSplitter()).keyBy(1).countWindow(3).sum(0).print();env.execute("flink window example");}
}

基于事件數量滑動窗口

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//基于事件數量滑動窗口data.flatMap(new DataSplitter()).keyBy(1).countWindow(4, 3).sum(0).print();*env.execute("flink window example");}
}

基于會話時間窗口

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//基于會話時間窗口data.flatMap(new DataSplitter()).keyBy(v->v.f0).window(ProcessingTimeSessionWindows.withGap(Time.seconds(5)))//表示如果 5s 內沒出現數據則認為超出會話時長，然后計算這個窗口的和.sum(1).print();env.execute("flink window example");}
}

滾動窗口（Tumbling Window）

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");//滾動窗口（Tumbling Window） 基于處理時間的 30 秒滾動窗口data.flatMap(new DataSplitter()).keyBy(v->v.f0).window(TumblingProcessingTimeWindows.of(Time.seconds(30))).sum(1).print();;env.execute("flink window example");}
}

基于事件時間

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");// 基于事件時間的 30 秒滾動窗口data.flatMap(new DataSplitter()).keyBy(v->v.f0).assignTimestampsAndWatermarks(/* 分配時間戳和水印 */).window(TumblingEventTimeWindows.of(Time.seconds(30))).sum(1).print();env.execute("flink window example");}
}

滑動窗口（Sliding Window）

基于處理時間

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");// 基于處理時間的 30 秒滑動窗口，滑動間隔為 10 秒data.flatMap(new DataSplitter()).keyBy(v->v.f0).window(SlidingProcessingTimeWindows.of(Time.seconds(30), Time.seconds(10))).sum(1).print();env.execute("flink window example");}
}

基于事件時間

/*** Desc: Flink Window 學習*/
@Slf4j
public class WindowsDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();String brokers = "localhost:9092";KafkaSource<String> source = KafkaSource.<String>builder().setBootstrapServers(brokers).setTopics(TOPIC_NAME).setGroupId("my-group").setStartingOffsets(OffsetsInitializer.earliest()).setValueOnlyDeserializer(new SimpleStringSchema()).build();DataStreamSource<String> data = env.fromSource(source, WatermarkStrategy.noWatermarks(), "wfgxxx");// 基于事件時間的 30 秒滑動窗口，滑動間隔為 10 秒  data.flatMap(new DataSplitter()).keyBy(v->v.f0).assignTimestampsAndWatermarks(/* 分配時間戳和水印 */).window(SlidingEventTimeWindows.of(Time.seconds(30), Time.seconds(10))).sum(1).print();env.execute("flink window example");}
}