scala Wordcount

package my.bigdata.scala08

import scala.collection.mutable
import scala.collection.mutable.ArrayBuffer
import scala.io.Source

/** scala word count
  * Created by lq on 2017/8/7.
  */
object Task2 {

  /**
    * basic 核心是外部变量 + map.getOrElse
    */
  def scalaWC0(): Unit ={
    val in = new java.util.Scanner(new java.io.File("myfile.txt"))
    //var tt = null;
    val words = new ArrayBuffer[String]
    while(in.hasNext())
    {
      words ++= in.next().split("\\s+")
    }
    var map = Map[String,Int]()
    for(key <- words) {
      map +=(key -> (map.getOrElse(key, 0)+1))
      println((key -> (map.getOrElse(key, 0)+1)))
      //
    }
    print(map)
  }

  /**
    * basic 核心是外部变量 + map.getOrElse
    */
  def scalaWC1(): Unit ={
    val lines = Source.fromFile("myfile.txt").getLines();
    val resMap = new mutable.HashMap[String,Long]()
    //lines.map(line=>{line.split("\\s+")}).
    for(line <- lines){
      val fields = line.split("\\s+")
      for(f <- fields){
        val v =  resMap.getOrElse(f,0L);
        resMap.put(f,v+1L)
      }
    }
    resMap.foreach(println(_))
  }

  /**
    * 利用遍历+外部变量map实现
    */
  def scalaWC5(): Unit = {
    val lines = Source.fromFile("myfile.txt").getLines();
    val resMap = new mutable.HashMap[String, Long]()
    val res = lines.flatMap(_.split("\\s+")).map(t=>{resMap+=((t,resMap.getOrElse(t,0L)+1L))})
    println(resMap)
  }

  /**
    * 利用遍历+外部变量map实现
    */
  def scalaWC4(): Unit = {
    val lines = Source.fromFile("myfile.txt").getLines();
    val resMap = new mutable.HashMap[String, Long]()
    val res = lines.flatMap(line => {
      line.split("\\s+")}).foldLeft(resMap)((x,y:String)=>{
      resMap+=((y,resMap.getOrElse(y,0L)+1L))
    })
    println(resMap)
  }

  /**
    * advance 没有用到外部变量, foldLeft 传入一个map,然后传入一个偏函数,偏函数结合遍历的数据处理map返回一个map,最后整个函数返回一个map
    */
  def scalaWC41(): Unit = {
    val lines = Source.fromFile("myfile.txt").getLines();
    val res = lines.flatMap(_.split("\\s+")).foldLeft(mutable.Map[String,Long]())((m,y:String)=>{
      m += ((y,m.getOrElse(y,0L)+1L))
    })
    println(res)
  }

  /**
    * advance 没有用到外部变量, foldLeft的遍历所有的功能
    */
  def scalaWC2(): Unit ={
    val lines = Source.fromFile("myfile.txt").getLines();
    //val resMap = new mutable.HashMap[String,Long]()
    val res = lines.map(line=>{line.split("\\s+").toBuffer})
      .toList.flatMap(x=>x).map((_,1)).groupBy(_._1)
      .map(x=>(x._1,x._2.foldLeft(0)((sum,t) =>{sum + t._2})))
    println(res)
  }

  /**
    * advance 没有用到外部变量, reduceLeft的遍历所有的功能
    */
  def scalaWC3(): Unit ={
    val lines = Source.fromFile("myfile.txt").getLines();
    val res = lines.map(line=>{line.split("\\s+").toBuffer})
      .toList.flatMap(x=>x).map((_,1)).groupBy(_._1)
      .map(x=>(x._2.reduceLeft((x,y)=>{
        (x._1,x._2+y._2)
      })))
    //出现
    println(res)
  }

  /**
    * 主函数
    *
    * @param args
    */
  def main(args: Array[String]): Unit = {
    // arrFun2()
    // printJavaProp()
    //customMap()
    scalaWC41()
  }
  /*
  总结:wc的实现思路
  1.循环+外部map变量+map特性
  2.利用集合的方法,通过各种变换,的到结果
  3.集合具有遍历的方法有map,filter,foreach,reduceLeft,foldLeft,这些加上外部map变量+map特性都能实现WordCount

   */
}