先附上scala实现的代码！
package cn.spark.study.core
import org.apache.spark.SparkConf
import org.apache.spark.SparkContext
/**
* @author Administrator
*/
object WordCount {
def main(args: Array[String]) {
val conf = new SparkConf()
.setAppName("WordCount");
val sc = new SparkContext(conf)
val lines = sc.textFile("hdfs://spark1:9000/spark.txt", 1);
val words = lines.flatMap { line => line.split(" ") }
val pairs = words.map { word => (word, 1) }
val wordCounts = pairs.reduceByKey { _ + _ }
wordCounts.foreach(wordCount => println(wordCount._1 + " appeared " + wordCount._2 + " times."))
}
}
下面的图就是根据上面的代码一步一步剖析底层实现的原理！
从下面的箭头看整个逻辑流程！
