Flink（五） —— DataStream API

Source

从自定义的集合中读取数据

/**
    * 从集合中读取数据
    */
  def readDataFromCollection(): Unit = {

    val env = StreamExecutionEnvironment.getExecutionEnvironment

    // 1.从自定义的集合中读取数据
    val list = List(
      SensorReading("sensor1", 153242, 35.8),
      SensorReading("sensor2", 153222, 15.4),
      SensorReading("sensor3", 153142, 6.7),
      SensorReading("sensor4", 151242, 38.7))

    val stream1 = env.fromCollection(list)

    stream1.print("stream1").setParallelism(1)

    env.execute("source test")
  }

从Kafka中读取数据

引入依赖

    <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-connector-kafka-0.10_2.11</artifactId>
            <version>1.7.2</version>
        </dependency>

代码

  /**
    * 从kafka中读取数据
    */
  def readDataFromKafka(): Unit = {

    val env = StreamExecutionEnvironment.getExecutionEnvironment

    val props = new Properties()
    props.setProperty("bootstrap.servers", "localhost:9092")
    props.setProperty("key.serializer", "org.apache.kafka.common.serialization.StringSerializer")
    props.setProperty("value.serializer", "org.apache.kafka.common.serialization.StringSerializer")
    props.setProperty("group.id", "flink-demo")
    props.setProperty("auto.offset.reset", "latest")

    val stream1 = env.addSource(new FlinkKafkaConsumer010[String]("flinkdemo",new SimpleStringSchema(),props))

    stream1.print("stream1").setParallelism(1)

    env.execute("source test")
  }

从自定义的Source中读取数据

class SensorSource() extends SourceFunction[SensorReading] {

  var running: Boolean = true

  // 取消数据源的生成
  override def cancel(): Unit = {
    running = false
  }

  // 生成数据
  override def run(sourceContext: SourceContext[SensorReading]): Unit = {
    // 初始化一个随机数发生器
    val rand = new Random()

    var curTemp = 1.to(10).map(
      i => ("sensor_" + i, 60 + rand.nextGaussian() * 20)
    )

    while (running) {

      curTemp = curTemp.map(
        t => (t._1, t._2 + rand.nextGaussian())
      )

      val curTime = System.currentTimeMillis()

      curTemp.foreach(
        t => sourceContext.collect(SensorReading(t._1, curTime, t._2))
      )

      Thread.sleep(500)

    }
  }
}

Transform

样例数据

senor_1,1,10
senor_2,2,20
senor_3,3,40
senor_4,4,30
senor_5,5,30
senor_6,6,60
senor_1,7,70

map、reduce、keyBy

map

• DataStream -> DataStream
• 通过应用给定的函数，对原先DataStream中的每个元素进行处理，获得一个新的DataStream

keyBy

• DataStream -> KeyedStream[T,JavaTuple]
• 对DataStream中的元素按照给定的表达式进行分组

reduce

• KeyedStream -> DataStream
• 通过规约原有DataStream中的元素，返回一个新的DataStream

  /**
    * 使用map、reduce
    */
  def testMap(): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(1)

    val streamFromFile = env.readTextFile("senor.txt")
    val dataStream: DataStream[SensorReading] = streamFromFile.map(data => {
      val dataArray = data.split(",")
      SensorReading(dataArray(0).trim, dataArray(1).toLong, dataArray(2).trim.toDouble)
    })
      .keyBy("id")
      .reduce((x, y) => {
        SensorReading(x.id, x.timestamp + 1, y.temperature + x.temperature)
      })

    dataStream.print()

    env.execute()
  }

split、select

split

• DataStream → SplitStream
• 按照指定标准将指定的DataStream拆分成多个流用SplitStream来表示

select

• SplitStream → DataStream
• 跟split搭配使用，从SplitStream中选择一个或多个流

def testSplit(): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(1)

    val streamFromFile = env.readTextFile("senor.txt")
    val dataStream: DataStream[SensorReading] = streamFromFile.map(data => {
      val dataArray = data.split(",")
      SensorReading(dataArray(0).trim, dataArray(1).toLong, dataArray(2).trim.toDouble)
    })

    // 多流转换算子
    val splitStream = dataStream.split(data => {
      if (data.temperature > 20) Seq("high") else Seq("low")
    })

    val high = splitStream.select("high")
    val low = splitStream.select("low")
    val all = splitStream.select("high", "low")

    high.print("high")
    low.print("low")
    all.print("all")

    env.execute()
  }

connect、coMap、coFlatMap

connect

• DataStream,DataStream -> ConnectedStreams

coMap

• ConnectedStreams -> DataStream

def testConnect(): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(1)

    val streamFromFile = env.readTextFile("senor.txt")
    val dataStream: DataStream[SensorReading] = streamFromFile.map(data => {
      val dataArray = data.split(",")
      SensorReading(dataArray(0).trim, dataArray(1).toLong, dataArray(2).trim.toDouble)
    })

    // 多流转换算子
    val splitStream = dataStream.split(data => {
      if (data.temperature > 20) Seq("high") else Seq("low")
    })

    val high = splitStream.select("high")
    val low = splitStream.select("low")

    // 创建一个新的数据流，数据类型与high、low不同
    val warning = high.map(data => (data.id, data.temperature))
    // 得到ConnectedStreams[T, T2]
    val connectedStreams = warning.connect(low)
    val coMapDataStreams = connectedStreams.map(data1 => (data1._1, data1._2, "warning"), data2 => (data2.temperature, "health"))

    coMapDataStreams.print()

    env.execute()
  }

UDF函数

Filter

def testFilter(): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(1)

    val streamFromFile = env.readTextFile("senor.txt")
    val dataStream: DataStream[SensorReading] = streamFromFile.map(data => {
      val dataArray = data.split(",")
      SensorReading(dataArray(0).trim, dataArray(1).toLong, dataArray(2).trim.toDouble)
    })

    dataStream.filter(new MyFilter()).print()

    env.execute()
  }

class MyFilter() extends FilterFunction[SensorReading] {
    override def filter(value: SensorReading): Boolean = {
      return value.id.startsWith("senor_1")
    }
  }

Sink

  def testFlinkSink2Kafka(): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    env.setParallelism(1)

    val streamFromFile = env.readTextFile("senor.txt")

    // Transform操作
    val dataStream = streamFromFile.map(data => {
      val dataArray = data.split(",")
      SensorReading(dataArray(0).trim, dataArray(1).toLong, dataArray(2).trim.toDouble).toString
    })

    // sink
    dataStream.addSink(new FlinkKafkaProducer010[String]("localhost:9092", "sinkTest", new SimpleStringSchema()))

    env.execute()
  }

参考文档

Basic API Concepts

Flink算子使用方法及实例演示：union和connect