当前位置: 首页 > news >正文

【Flink-scala】DataStream编程模型之延迟数据处理

DataStream API编程模型

1.【Flink-Scala】DataStream编程模型之数据源、数据转换、数据输出
2.【Flink-scala】DataStream编程模型之 窗口的划分-时间概念-窗口计算程序
3.【Flink-scala】DataStream编程模型之水位线
4.【Flink-scala】DataStream编程模型之窗口计算-触发器-驱逐器


文章目录

  • DataStream API编程模型
  • 一、延迟数据处理
    • 1.1 侧输出
    • 1.2代码实例
      • 1.2.1 代码运行结果
      • 1.2.2 结果分析
    • 1.3代码运行展示
    • 1.4 题外话


一、延迟数据处理

前一小节已经讲了水位线的相关概念,默认情况下,当水位线超过窗口结束时间之后,再有之前的数据到达时,这些数据会被删除。

为了避免有些迟到的数据被删除,因此产生了allowedLateness的概念。

allowedLateness就是针对事件时间而言,对于水位线超过窗口结束时间之后,还允许有一段时间(也是以事件时间来衡量)来等待之前的数据到达,以便再次处理这些数据。

对于窗口计算而言,如果没有设置allowedLateness,窗口触发计算以后就会被销毁

设置了allowedLateness以后,只有水位线大于“窗口结束时间+allowedLateness”时,窗口才会被销毁。

当没有指定allowedLateness,默认值为0.

1.1 侧输出

通常情况下,用户虽然希望对迟到的数据进行窗口计算,但并不想将结果混入正常的计算流程中,而是想将延迟数据和结果保存到数据库中,便于后期对延时数据进行分析。对这种情况,就需要借助于“侧输出”(Side Output)来处理

用sideOutputLateData(OutputTag)来标记迟到数据计算的结果,然后再使用getSideOutput(lateOutputTag)从窗口中获取lateOutputTag标签对应的数据,之后转成独立的DataStream数据集进行处理

1.2代码实例

如下:

import java.text.SimpleDateFormat
import org.apache.flink.api.common.eventtime.{SerializableTimestampAssigner, TimestampAssigner, TimestampAssignerSupplier, Watermark, WatermarkGenerator, WatermarkGeneratorSupplier, WatermarkOutput, WatermarkStrategy}
import org.apache.flink.streaming.api.scala._
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.scala.StreamExecutionEnvironment
import org.apache.flink.streaming.api.windowing.assigners.TumblingEventTimeWindows
import org.apache.flink.streaming.api.windowing.time.Timecase class StockPrice(stockId:String,timeStamp:Long,price:Double)object AllowedLatenessTest {def main(args: Array[String]): Unit = {//设定执行环境val env = StreamExecutionEnvironment.getExecutionEnvironment//设定时间特性为事件时间env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)//设定程序并行度env.setParallelism(1)//创建数据源val source = env.socketTextStream("localhost", 9999)//指定针对数据流的转换操作逻辑val stockDataStream = source.map(s => s.split(",")).map(s=>StockPrice(s(0).toString,s(1).toLong,s(2).toDouble))//为数据流分配时间戳和水位线val watermarkDataStream = stockDataStream.assignTimestampsAndWatermarks(new MyWatermarkStrategy)//执行窗口计算val lateData = new OutputTag[StockPrice]("late")val sumStream = watermarkDataStream.keyBy("stockId").window(TumblingEventTimeWindows.of(Time.seconds(3))).allowedLateness(Time.seconds(2L))//注意这里.sideOutputLateData(lateData).reduce((s1, s2) => StockPrice(s1.stockId,s1.timeStamp, s1.price + s2.price))//打印输出sumStream.print("window计算结果:")val late = sumStream.getSideOutput(lateData)late.print("迟到的数据:")//指定名称并触发流计算env.execute("AllowedLatenessTest")}//指定水位线生成策略class MyWatermarkStrategy extends WatermarkStrategy[StockPrice] {override def createTimestampAssigner(context:TimestampAssignerSupplier.Context):TimestampAssigner[StockPrice]={new SerializableTimestampAssigner[StockPrice] {override def extractTimestamp(element: StockPrice, recordTimestamp: Long): Long = {element.timeStamp //从到达消息中提取时间戳}}}override def createWatermarkGenerator(context:WatermarkGeneratorSupplier.Context): WatermarkGenerator[StockPrice] ={new WatermarkGenerator[StockPrice](){val maxOutOfOrderness = 10000L //设定最大延迟为10秒var currentMaxTimestamp: Long = 0Lvar a: Watermark = nullval format = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss.SSS")override def onEvent(element: StockPrice, eventTimestamp: Long, output:WatermarkOutput): Unit = {currentMaxTimestamp = Math.max(eventTimestamp, currentMaxTimestamp)a = new Watermark(currentMaxTimestamp - maxOutOfOrderness)output.emitWatermark(a)println("timestamp:" + element.stockId + "," + element.timeStamp + "|" + format.format(element.timeStamp) + "," + currentMaxTimestamp + "|" + format.format(currentMaxTimestamp) + "," + a.toString)}
override def onPeriodicEmit(output:WatermarkOutput): Unit = {// 没有使用周期性发送水印,因此这里没有执行任何操作}}}}
}

先注意这里:

  val watermarkDataStream = stockDataStream.assignTimestampsAndWatermarks(new MyWatermarkStrategy)//执行窗口计算val lateData = new OutputTag[StockPrice]("late")val sumStream = watermarkDataStream.keyBy("stockId").window(TumblingEventTimeWindows.of(Time.seconds(3))).allowedLateness(Time.seconds(2L))   //注意这里.sideOutputLateData(lateData).reduce((s1, s2) => StockPrice(s1.stockId,s1.timeStamp, s1.price + s2.price))

允许元素延迟到达最多 2 秒。即,如果一个元素的时间戳在窗口结束后的 2 秒内到达,它仍然会被包含在窗口的计算中。

 .allowedLateness(Time.seconds(2L))  设置时间为2s。

其余的就是生成水位线和时间戳了,这就不用解释啦。

1.2.1 代码运行结果

1.在nc端输入
stock_1,1602031567000,8.14
stock_1,1602031571000,8.23
stock_1,1602031577000,8.24
stock_1,1602031578000,8.87
stock_1,1602031579000,8.55
stock_1,1602031577000,8.24
stock_1,1602031581000,8.43
stock_1,1602031582000,8.78
stock_1,1602031581000,8.76
stock_1,1602031579000,8.55
stock_1,1602031591000,8.13
stock_1,1602031581000,8.34
stock_1,1602031580000,8.45
stock_1,1602031579000,8.33
stock_1,1602031578000,8.56
stock_1,1602031577000,8.32
然后启动程序

在这里插入图片描述
分析:水位线生成策略:

每次收到一个新的事件,都会比较当前事件的时间戳和之前记录的最大时间戳,更新 currentMaxTimestamp。
水位线被设定为 currentMaxTimestamp - maxOutOfOrderness,即允许最大 10 秒的延迟

再来看迟到数据:

输入数据是:
stock_1,1602031567000,8.14
stock_1,1602031571000,8.23
stock_1,1602031577000,8.24
stock_1,1602031578000,8.87
stock_1,1602031579000,8.55
stock_1,1602031577000,8.24
stock_1,1602031581000,8.43
stock_1,1602031582000,8.78
stock_1,1602031581000,8.76
stock_1,1602031579000,8.55
stock_1,1602031591000,8.13
stock_1,1602031581000,8.34
stock_1,1602031580000,8.45
stock_1,1602031579000,8.33
stock_1,1602031578000,8.56
stock_1,1602031577000,8.32

数据超出窗口结算数据后2s到达算迟到,任何时间戳比水位线晚 10 秒的事件都会被视为迟到。
之前我看书上写的有点疑问,然后拥有环境的网站也有点问题,我就自己本地了,windows 下载了个netcat,运行结果如下:
在这里插入图片描述

timestamp:stock_1,1602031567000|2020-10-07 08:46:07.000,1602031567000|2020-10-07 08:46:07.000,Watermark @ 1602031557000 (2020-10-07 08:45:57.000)
timestamp:stock_1,1602031571000|2020-10-07 08:46:11.000,1602031571000|2020-10-07 08:46:11.000,Watermark @ 1602031561000 (2020-10-07 08:46:01.000)
timestamp:stock_1,1602031577000|2020-10-07 08:46:17.000,1602031577000|2020-10-07 08:46:17.000,Watermark @ 1602031567000 (2020-10-07 08:46:07.000)
timestamp:stock_1,1602031578000|2020-10-07 08:46:18.000,1602031578000|2020-10-07 08:46:18.000,Watermark @ 1602031568000 (2020-10-07 08:46:08.000)
timestamp:stock_1,1602031579000|2020-10-07 08:46:19.000,1602031579000|2020-10-07 08:46:19.000,Watermark @ 1602031569000 (2020-10-07 08:46:09.000)
timestamp:stock_1,1602031577000|2020-10-07 08:46:17.000,1602031579000|2020-10-07 08:46:19.000,Watermark @ 1602031569000 (2020-10-07 08:46:09.000)
timestamp:stock_1,1602031581000|2020-10-07 08:46:21.000,1602031581000|2020-10-07 08:46:21.000,Watermark @ 1602031571000 (2020-10-07 08:46:11.000)
timestamp:stock_1,1602031582000|2020-10-07 08:46:22.000,1602031582000|2020-10-07 08:46:22.000,Watermark @ 1602031572000 (2020-10-07 08:46:12.000)
timestamp:stock_1,1602031581000|2020-10-07 08:46:21.000,1602031582000|2020-10-07 08:46:22.000,Watermark @ 1602031572000 (2020-10-07 08:46:12.000)
timestamp:stock_1,1602031579000|2020-10-07 08:46:19.000,1602031582000|2020-10-07 08:46:22.000,Watermark @ 1602031572000 (2020-10-07 08:46:12.000)
timestamp:stock_1,1602031591000|2020-10-07 08:46:31.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
timestamp:stock_1,1602031581000|2020-10-07 08:46:21.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
timestamp:stock_1,1602031580000|2020-10-07 08:46:20.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
timestamp:stock_1,1602031579000|2020-10-07 08:46:19.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
timestamp:stock_1,1602031578000|2020-10-07 08:46:18.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
window计算结果:> StockPrice(stock_1,1602031567000,8.14)
window计算结果:> StockPrice(stock_1,1602031571000,8.23)
window计算结果:> StockPrice(stock_1,1602031577000,16.48)
window计算结果:> StockPrice(stock_1,1602031578000,25.970000000000002)
window计算结果:> StockPrice(stock_1,1602031578000,34.42)
window计算结果:> StockPrice(stock_1,1602031578000,42.75)
window计算结果:> StockPrice(stock_1,1602031578000,51.31)
timestamp:stock_1,1602031577000|2020-10-07 08:46:17.000,1602031591000|2020-10-07 08:46:31.000,Watermark @ 1602031581000 (2020-10-07 08:46:21.000)
迟到的数据:> StockPrice(stock_1,1602031577000,8.32)

这里是windows运行结果,可能稍微和ubuntu不一样

1.2.2 结果分析

  • stock_1,1602031567000,8.14
    水位线 557
    窗口时间**[567,570)**
    当前事件:stock_1,1602031567000,8.14
window计算结果:> StockPrice(stock_1,1602031567000,8.14)
  • stock_1,1602031571000,8.23
    水位线: 561
    窗口时间:[571,573)
    当前窗口事件 stock_1,1602031571000,8.23
window计算结果:> StockPrice(stock_1,1602031571000,8.23)
  • stock_1,1602031577000,8.24
    水位线:567
    窗口时间:[577,580)
    当前窗口事件:stock_1,1602031577000,8.24
    8.24+8.23
window计算结果:> StockPrice(stock_1,1602031577000,16.48)
  • stock_1,1602031578000,8.87
    水位线:568
    窗口时间:和上一个相同 [577,580)
    当前窗口事件:stock_1,1602031578000,8.87
    stock_1,1602031577000,8.24
    16.48+8.87
window计算结果:> StockPrice(stock_1,1602031578000,25.970000000000002)
  • stock_1,1602031579000,8.55
    水位线:569
    窗口时间:和上一个相同 [577,580)
    当前窗口事件(3个)
    stock_1,1602031578000,8.87
    stock_1,1602031577000,8.24

    stock_1,1602031579000,8.55
    25.97+8.55
window计算结果:> StockPrice(stock_1,1602031578000,34.42)
  • stock_1,1602031577000,8.24
    水位线:这里-10s为567,最大水位线为569,那么最大水位线569
    窗口时间: 属于[567,570) 没有迟到
    当前窗口事件:
    stock_1,1602031567000,8.14
    stock_1,1602031577000,8.24
    34.42+8.24
window计算结果:> StockPrice(stock_1,1602031578000,42.75)
  • stock_1,1602031581000,8.43
    水位线:571
    窗口时间:[581,584)
    当前窗口事件 stock_1,1602031581000,8.43
    注意 571水位线,第一个事件事件窗口时间[567,570),这个就应该结束计算了、
window计算结果:> StockPrice(stock_1,1602031578000,51.31)
之后就没了
为什么没了呢,窗口还没计算完呢,一直在监听啊。
水位线也不涨了
  • stock_1,1602031582000,8.78
    水位线:572
    窗口时间:[581,584)
    当前窗口事件:
    stock_1,1602031581000,8.43
    stock_1,1602031582000,8.78

  • stock_1,1602031581000,8.76
    水位线:572
    窗口时间:[581,584)
    当前窗口事件:
    stock_1,1602031581000,8.43
    stock_1,1602031582000,8.78

    stock_1,1602031581000,8.76

  • stock_1,1602031579000,8.55
    水位线:572
    窗口时间:[577,580)
    当前窗口事件:(4个)
    stock_1,1602031578000,8.87
    stock_1,1602031577000,8.24
    stock_1,1602031579000,8.55

    stock_1,1602031579000,8.55

  • stock_1,1602031591000,8.13
    水位线:581
    窗口时间:[591,594)
    当前窗口事件:
    stock_1,1602031591000,8.13

  • stock_1,1602031581000,8.34
    -水位线:581
    窗口时间:[581,584)
    (此时:[577,580)就应该结束了)
    窗口事件:
    stock_1,1602031581000,8.43
    stock_1,1602031582000,8.78
    stock_1,1602031581000,8.76

    stock_1,1602031581000,8.34

  • stock_1,1602031580000,8.45
    水位线:581
    窗口时间 [580,583)
    窗口事件:stock_1,1602031580000,8.45

  • stock_1,1602031579000,8.33
    水位线:581
    窗口时间:[577,580) 水位线在581的时候就结束了,但它在 allowedLateness 的 2 秒内到达,因此它 不会被视为迟到数据。

  • stock_1,1602031578000,8.56
    水位线:581
    窗口时间属于 [577,580)
    判断水位线和时间戳的关系:水位线超过窗口结束时间
    事件是否迟到:允许超出2s,正好就在580的边界,根据 Flink 的延迟数据规则,任何事件的时间戳如果大于当前窗口结束时间,并且超出了 allowedLateness(即 2 秒),就会被视为迟到数据。这个事件 stock_1,1602031578000,8.56 的时间戳刚好等于窗口 [577, 580) 的结束时间,并且它的时间戳在当前窗口结束后刚好到达,所以 它不会被视为迟到数据。

找了chatgpt问了一下:

这个事件不算迟到数据。尽管它的时间戳正好是窗口结束的边界,但它仍然属于这个窗口 [577, 580) 的一部分,不超出窗口时间范围,也没有超过 allowedLateness 设置的 2 秒,所以它被处理为正常数据。
为什么不算迟到?
Flink 的窗口操作是基于事件时间戳的,窗口 [577, 580) 的结束时间是 1602031578000。当水位线达到 581 时,窗口已经结束,但如果事件的时间戳恰好是窗口的结束时间,它依然被认为是属于该窗口的。
迟到数据的定义:迟到数据是指事件的时间戳超过了当前窗口结束时间,并且超出了允许的迟到时间。在这种情况下,事件时间戳与窗口结束时间对齐,因此并不算迟到。

  • stock_1,1602031577000,8.32
    水位线 581
    窗口时间:[577,580)
    577+2<581
    当前事件的时间戳 1602031577000 恰好等于窗口的起始时间 577,但水位线已经推进到了 581。
    窗口已经结束:水位线达到 581 时,窗口 [577, 580) 的数据已经完全处理完毕,因此任何时间戳在 580 之后到达的事件,都被视为迟到数据。

  • 说明:其中的删除线,我是手动分析的他们如何迟到,但是后来代码发现,数据是一个个来的,你没有超出水位线,那么我就立马把你这条数据计算了,因此输出结果就是累加起来的。

1.3代码运行展示

本来使用的是头歌平台提供的环境,但是环境有问题了,自己手动搭建了一下。幸亏是socket,windows还能实现监听自己,就把代码放在这里啦!
pom.xml

<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd"><modelVersion>4.0.0</modelVersion><groupId>org.example</groupId><artifactId>Flink_scala2.13</artifactId><version>1.0-SNAPSHOT</version><properties><maven.compiler.source>8</maven.compiler.source><maven.compiler.target>8</maven.compiler.target><flink.version>1.11.0</flink.version> <!-- 使用 Flink 1.15.0 --><target.java.version>1.8</target.java.version><scala.binary.version>2.12</scala.binary.version><scala.version>2.12.16</scala.version></properties><dependencies><!-- 引入 Flink 相关依赖--><dependency><groupId>org.apache.flink</groupId><artifactId>flink-scala_${scala.binary.version}</artifactId><version>${flink.version}</version><!--			<scope>provided</scope>--></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-streaming-scala_${scala.binary.version}</artifactId><version>${flink.version}</version><!--			<scope>provided</scope>--></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-clients_${scala.binary.version}</artifactId><version>${flink.version}</version><!--			<scope>provided</scope>--></dependency></dependencies>
</project>

test1.scala (AllowedLatenessTest函数)


/*
stock_1,1602031567000,8.14
stock_1,1602031571000,8.23
stock_1,1602031577000,8.24
stock_1,1602031578000,8.87
stock_1,1602031579000,8.55
stock_1,1602031577000,8.24
stock_1,1602031581000,8.43
stock_1,1602031582000,8.78
stock_1,1602031581000,8.76
stock_1,1602031579000,8.55
stock_1,1602031591000,8.13
stock_1,1602031581000,8.34
stock_1,1602031580000,8.45
stock_1,1602031579000,8.33
stock_1,1602031578000,8.56
stock_1,1602031577000,8.32*/
import java.text.SimpleDateFormat
import org.apache.flink.api.common.eventtime.{SerializableTimestampAssigner, TimestampAssigner, TimestampAssignerSupplier, Watermark, WatermarkGenerator, WatermarkGeneratorSupplier, WatermarkOutput, WatermarkStrategy}
import org.apache.flink.streaming.api.scala._
import org.apache.flink.streaming.api.TimeCharacteristic
import org.apache.flink.streaming.api.scala.StreamExecutionEnvironment
import org.apache.flink.streaming.api.windowing.assigners.TumblingEventTimeWindows
import org.apache.flink.streaming.api.windowing.time.Time
case class StockPrice(stockId:String,timeStamp:Long,price:Double)object test1 {def main(args: Array[String]): Unit = {// *************************** Begin ****************************//设定执行环境val env = StreamExecutionEnvironment.getExecutionEnvironment//设定时间特性为事件时间env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime)//设定程序并行度env.setParallelism(1)//创建数据源val source = env.socketTextStream("localhost", 9999)//指定针对数据流的转换操作逻辑val stockDataStream = source.map(s => s.split(",")).map(s=>StockPrice(s(0).toString,s(1).toLong,s(2).toDouble))//为数据流分配时间戳和水位线val watermarkDataStream = stockDataStream.assignTimestampsAndWatermarks(new MyWatermarkStrategy)//执行窗口计算val lateData = new OutputTag[StockPrice]("late")val sumStream = watermarkDataStream.keyBy("stockId").window(TumblingEventTimeWindows.of(Time.seconds(3))).allowedLateness(Time.seconds(2L)).sideOutputLateData(lateData).reduce((s1, s2) => StockPrice(s1.stockId,s1.timeStamp, s1.price + s2.price))// **************************** End *****************************//打印输出sumStream.print("window计算结果:")val late = sumStream.getSideOutput(lateData)late.print("迟到的数据:")//指定名称并触发流计算env.execute("AllowedLatenessTest")}//指定水位线生成策略class MyWatermarkStrategy extends WatermarkStrategy[StockPrice] {override def createTimestampAssigner(context:TimestampAssignerSupplier.Context):TimestampAssigner[StockPrice]={new SerializableTimestampAssigner[StockPrice] {override def extractTimestamp(element: StockPrice, recordTimestamp: Long): Long = {element.timeStamp //从到达消息中提取时间戳}}}override def createWatermarkGenerator(context:WatermarkGeneratorSupplier.Context): WatermarkGenerator[StockPrice] ={new WatermarkGenerator[StockPrice](){val maxOutOfOrderness = 10000L //设定最大延迟为10秒var currentMaxTimestamp: Long = 0Lvar a: Watermark = nullval format = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss.SSS")override def onEvent(element: StockPrice, eventTimestamp: Long, output:WatermarkOutput): Unit = {currentMaxTimestamp = Math.max(eventTimestamp, currentMaxTimestamp)a = new Watermark(currentMaxTimestamp - maxOutOfOrderness)output.emitWatermark(a)println("timestamp:" + element.stockId + "," + element.timeStamp + "|" + format.format(element.timeStamp) + "," + currentMaxTimestamp + "|" + format.format(currentMaxTimestamp) + "," + a.toString)}override def onPeriodicEmit(output:WatermarkOutput): Unit = {// 没有使用周期性发送水印,因此这里没有执行任何操作}}}}
}

这个是头歌提供的pom.xml,二者应该都能使用,但是你需要下载flink和scala,这不再多说,版本还要匹配哦

<project><groupId>cn.edu.xmu.dblab</groupId><artifactId>wordcount_myID</artifactId><modelVersion>4.0.0</modelVersion><name>WordCount</name><packaging>jar</packaging><version>1.0</version><repositories><repository><id>alimaven</id><name>aliyun maven</name><url>http://maven.aliyun.com/nexus/content/groups/public/</url></repository></repositories><properties><maven.compiler.source>1.8</maven.compiler.source><maven.compiler.target>1.8</maven.compiler.target><encoding>UTF-8</encoding><scala.version>2.12.2</scala.version><scala.binary.version>2.12</scala.binary.version><hadoop.version>2.7.7</hadoop.version><flink.version>1.11.2</flink.version></properties><dependencies><dependency><groupId>org.scala-lang</groupId><artifactId>scala-library</artifactId><version>${scala.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-java</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-streaming-java_${scala.binary.version}</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-scala_${scala.binary.version}</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-streaming-scala_${scala.binary.version}</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-clients_${scala.binary.version}</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-connector-kafka-0.10_${scala.binary.version}</artifactId><version>${flink.version}</version></dependency><dependency><groupId>org.apache.hadoop</groupId><artifactId>hadoop-client</artifactId><version>${hadoop.version}</version></dependency><dependency><groupId>mysql</groupId><artifactId>mysql-connector-java</artifactId><version>5.1.38</version></dependency><dependency><groupId>com.alibaba</groupId><artifactId>fastjson</artifactId><version>1.2.22</version></dependency><dependency><groupId>org.apache.flink</groupId><artifactId>flink-connector-redis_2.11</artifactId><version>1.1.5</version></dependency><dependency><groupId>redis.clients</groupId><artifactId>jedis</artifactId><version>2.8.0</version><scope>compile</scope></dependency></dependencies>
<build><plugins><plugin><groupId>net.alchim31.maven</groupId><artifactId>scala-maven-plugin</artifactId><version>3.2.0</version><executions><execution><goals><goal>compile</goal></goals></execution></executions></plugin><plugin><groupId>org.apache.maven.plugins</groupId><artifactId>maven-assembly-plugin</artifactId><version>3.0.0</version><configuration><descriptorRefs><descriptorRef>jar-with-dependencies</descriptorRef></descriptorRefs></configuration><executions><execution><id>make-assembly</id><phase>package</phase><goals><goal>single</goal></goals></execution></executions></plugin></plugins></build>
</project>

1.4 题外话

我使用的课本是厦门大学林子雨老师编写的《Flink编程基础》,当前我用的书是2021年9月第一版,第5.6章节的延迟数据处理中(201页),最后迟到的数据8.43写错了,应该是:8.32。正是书上写错了,我才有点疑问发现此处的问题。


http://www.mrgr.cn/news/80257.html

相关文章:

  • vscode+msys2+clang+xmake c++开发环境搭建
  • 2024CAT开发测试 Web应用赛道
  • Nginx配置示例教程
  • ARM64平台实时linux操作系统xenomai4(EVL)构建安装简述
  • 使用rust语言创建python模块(pyo3+maturin)
  • 高等动力学中的正则变换
  • 在 macOS 下安装和使用 Clang Static Analyzer
  • 哈希【C++实现】
  • 2024年底-Sre面试回顾
  • 使用create-react-app创建工程时报错处理
  • 事件代理详解
  • JVM 面试题相关总结
  • 0.基础语法
  • Python面试常见问题及答案7
  • 使用Python实现两组数据纵向排序
  • 12.3【OS】
  • 5.字符串
  • 12.16周一F34-Day27打卡
  • 三七互娱Android面试题及参考答案
  • R-CNN