Spark Streaming（二）—— 创建DStream

Spark Streaming（二）—— 创建DStream 1、文件数据源案例

需求：读取hdfs上的Herry.txt文件，进行词频统计

package com.zch.spark.streaming
​
import org.apache.spark.SparkConf
import org.apache.spark.streaming.{Seconds, StreamingContext}
​

object sparkStreaming01_FileWordCount {
  def main(args: Array[String]): Unit = {
    //1.初始化 Spark 配置信息
    val sparkConf = new SparkConf().setMaster("local[*]")
      .setAppName("StreamWordCount")
    
    //2.初始化 SparkStreamingContext
    val ssc = new StreamingContext(sparkConf, Seconds(5))
    
    //3.监控文件夹创建 DStream
    val dirStream = ssc.textFileStream("hdfs://zhaohui01:8020/Herry.txt")
    
    //4.将每一行数据做切分，形成一个个单词
    val wordStreams = dirStream.flatMap(_.split("t"))
    
    //5.将单词映射成元组（word,1）
    val wordAndoneStreams = wordStreams.map((_, 1))
    
    //6.将相同的单词次数做统计
    val wordAndCountStreams = wordAndOneStreams.reduceByKey(_ + _)
    
    //7.打印
    wordAndCountStreams.print()
    
    //8.启动 SparkStreamingContext
    ssc.start()
    ssc.awaitTermination()
  }
​
}

2、创建DStream Queue

package com.zch.spark.streaming
​
import org.apache.spark.SparkConf
import org.apache.spark.rdd.RDD
import org.apache.spark.streaming.{Seconds, StreamingContext}
​
import scala.collection.mutable
​

object sparkStreaming02_Queue {
  def main(args: Array[String]): Unit = {
    val sparkConf = new SparkConf()
      .setMaster("local[*]")
      .setAppName("queue")
​
    val ssc = new StreamingContext(sparkConf, Seconds(3))
​
    // 创建RDD队列
    val rddQueue = new mutable.Queue[RDD[Int]]()
    // 创建QueueInputDStream
    val inputStream = ssc.queueStream(rddQueue, oneAtATime = false)
​
    val mappedStream = inputStream.map((_, 1))
    val reducedStream = mappedStream.reduceByKey(_ + _)
    reducedStream.print()
​
    ssc.start()
​
    for(i <- 1 to 10){
      rddQueue += ssc.sparkContext.makeRDD(1 to 300 , 10)
      Thread.sleep(2000)
    }
  }
}

3、自定义数据源

需要继承Receiver，并实现onStart和onStop方法来自定义数据源

class MyReceiver extends Receiver[String](StorageLevel.MEMORY_ONLY){
​
  private var flag = true
​
  override def onStart(): Unit = {
​
    new Thread (() => while (flag){
​
      val string ="采集的数据为："+ new Random().nextInt(10).toString
        store(string)
        Thread.sleep(500)
    }).start()
  }
​
  override def onStop(): Unit = {
​
    flag = false
  }
}

4、Kafka数据源（重点）

需求：通过SparkStreaming从Kafka读取数据

1.导入依赖

    org.apache.spark
    spark-streaming-kafka-0-10_2.12
    3.1.2

2.编写代码

package com.zch.spark.streaming
​
import org.apache.kafka.clients.consumer.{ConsumerConfig, ConsumerRecord}
import org.apache.spark.SparkConf
import org.apache.spark.streaming.dstream.InputDStream
import org.apache.spark.streaming.kafka010.{ConsumerStrategies, KafkaUtils, LocationStrategies}
import org.apache.spark.streaming.{Seconds, StreamingContext}
​

object sparkStreaming04_Kafka {
  def main(args: Array[String]): Unit = {
​
    val sparkConf = new SparkConf()
      .setMaster("local[*]")
      .setAppName("kafka")
​
    val ssc = new StreamingContext(sparkConf, Seconds(3))
​
    val kafkaParam: Map[String, String] = Map[String, String](
      ConsumerConfig.BOOTSTRAP_SERVERS_ConFIG ->
        "zhaohui01:9092,zhaohui02:9092,zhaohui03:9092",
      ConsumerConfig.GROUP_ID_ConFIG -> "__consumer",
      "key.deserializer" -> "org.apache.kafka.common.serialization.StringDeserializer",
      "value.deserializer" -> "org.apache.kafka.common.serialization.StringDeserializer",
​
    )
​
    val kafkaDS: InputDStream[ConsumerRecord[String, String]] = KafkaUtils.createDirectStream[String, String](
      ssc,
      LocationStrategies.PreferConsistent,
      ConsumerStrategies.Subscribe[String, String](Set("first01"), kafkaParam)
    )
​
    kafkaDS.map(_.value()).print()
​
    ssc.start()
    ssc.awaitTermination()
  }
​
}