Java_spark简单例子

import org.apache.spark.{SparkContext, SparkConf}

/**
 * Created by spark on 15-1-19.
 * 根据key对K-V类型的RDD进行排序获得新的RDD
 */
object SortByKey {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    import org.apache.spark.SparkContext._
    val a = sc.parallelize(List("dog","cat","owl","gnu","ant"))
    val b = sc.parallelize(1 to a.count().toInt)
    val c = a.zip(b)

    //asc
    c.sortByKey(true).collect().foreach(print)
    //desc
    c.sortByKey(false).collect().foreach(print)
  }
}
/**
 * Created by spark on 15-1-19.
 * RDD1.subtract(RDD2):返回一个新的RDD，内容是：RDD1中存在的，RDD2中不存在的
 */
object Subtract {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 10)
    val b = sc.parallelize(1 to 3)
    //45678910
    //a.subtract(b).collect().foreach(print)

    val c = sc.parallelize(1 to 10)
    val d = sc.parallelize(List(1,2,3,11))
    //
    c.subtract(d).collect().foreach(print)
  }
}
/**
 * Created by spark on 15-1-19.
 * RDD1.subtractByKey(RDD2):返回一个新的RDD，内容是：RDD1 key中存在的，RDD2 key中不存在的
 */
object SubtractByKey {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    import org.apache.spark.SparkContext._
    val a = sc.parallelize(List("dog","he","word","hello"))
    val b = a.keyBy(_.length)

    val c = sc.parallelize(List("cat","first","everyone"))
    val d = c.keyBy(_.length)
    //(2,he)(4,word)
    b.subtractByKey(d).collect().foreach(print)
  }
}
/**
 * Created by spark on 15-1-19.
 * sumApprox没有出现我希望的结果
 */
object SumAndSumApprox {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 1000000)
    val b = a.sum()
    val c = a.sumApprox(0L,0.9).getFinalValue()

    println(b + " *** " + c)
  }
}
/**
 * Created by spark on 15-1-19.
 * 取出RDD的前n个元素，以数组的形式返回
 */
object Take {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 1000000)

    //
    a.take(10).foreach(print)
  }
}
/**
 * Created by spark on 15-1-19.
 * 对RDD元素进行升序排序
 * 取出前n个元素并以数组的形式放回
 */
object TakeOrdered {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(List("ff","aa","dd","cc"))

    //aacc
    a.takeOrdered(2).foreach(print)
  }
}
/**
 * Created by spark on 15-1-19.
 * 数据取样
 */
object TakeSample {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 10000)

    /**
     * 9048
        5358
        5216
        7301
        6303
        6179
        6151
        5304
        8115
        3869
     */
    a.takeSample(true , 10 , 1).foreach(println)
  }
}
/**
 * Created by spark on 15-1-19.
 * debug 详情信息显示
 */
object ToDebugString {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 9)
    val b = sc.parallelize(1 to 3)
    val c = a.subtract(b)

    c.toDebugString
  }
}
/**
 * Created by spark on 15-1-19.
 * 获得前几个最大值
 */
object Top {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 1000)
    val c = a.top(10)

    /**
     *1000
      999
      998
      997
      996
      995
      994
      993
      992
      991
     */
    c.foreach(println)
  }
}

/**
 * Union == ++ 把两个RDD合并为一个新的RDD
 */
object Union {
  def main(args: Array[String]) {
    val conf = new SparkConf().setAppName("spark-demo").setMaster("local")
    val sc = new SparkContext(conf)

    //import org.apache.spark.SparkContext._
    val a = sc.parallelize(1 to 3)
    val b = sc.parallelize(3 to 5)
    val c = a.union(b)
    val d = a ++ b

    /**
     *123345
     */
    c.collect().foreach(print)

    /**
     *123345
     */
    d.collect().foreach(print)
  }
}

--Java

package com.demo.sparkWordCount;

import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaPairRDD;
import org.apache.spark.api.java.JavaRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.function.VoidFunction;

import akka.japi.Function;
import scala.Tuple2;

/*
 * Ming Z M LI
 * */
public class FunctionDemo {
    /*
     * create Context
     */
    public static JavaSparkContext createContext() {

        SparkConf sparkConf = new SparkConf().setAppName("FunctionDemo").setMaster("local[*]");

        JavaSparkContext ctx = new JavaSparkContext(sparkConf);

        return ctx;

    }

    public static void main(String[] args) {

        demo5();
    }

    /*
     * RDD1.subtract(RDD2):返回一个新的RDD，内容是：RDD1中存在的，RDD2中不存在的
     */
    public static void demo2() {

        JavaSparkContext ctx = createContext();
        List<String> list1 = new ArrayList<String>();
        list1.add("hello1");
        list1.add("hello2");
        list1.add("hello3");
        list1.add("hello4");

        List<String> list2 = new ArrayList<String>();
        list2.add("hello3");
        list2.add("hello4");
        list2.add("world5");
        list2.add("world6");

        JavaRDD<String> a = ctx.parallelize(list1);
        JavaRDD<String> b = ctx.parallelize(list2);

        a.subtract(b).foreach(new VoidFunction<String>() {
            public void call(String t) throws Exception {
                System.out.println(t.toString());
            }
        });
    }

    /**
     * Created by spark on 15-1-19. RDD1.subtractByKey(RDD2):返回一个新的RDD，内容是：RDD1
     * key中存在的，RDD2 key中不存在的 foreach 结果带key (4, bird) (5, hello) (3, cat) output
     * - (4,bird) (4,bird)
     */

    public static void demo3() {
        JavaSparkContext ctx = createContext();
        JavaRDD<String> a = ctx.parallelize(new ArrayList<String>(Arrays.asList("cat", "hello", "bird", "bird")));
        JavaRDD<String> b = ctx.parallelize(new ArrayList<String>(Arrays.asList("cat", "hello", "testing")));

        JavaPairRDD<Integer, String> c = a.keyBy(new org.apache.spark.api.java.function.Function<String, Integer>() {

            public Integer call(String v1) throws Exception {

                return v1.length();
            }

        });

        // c.foreach(new VoidFunction<Tuple2<Integer,String>>(){
        //
        // public void call(Tuple2<Integer, String> t) throws Exception {
        // // TODO Auto-generated method stub
        // System.out.println("("+t._1+", "+t._2+")");
        // }
        // });

        JavaPairRDD<Integer, String> d = b.keyBy(new org.apache.spark.api.java.function.Function<String, Integer>() {

            public Integer call(String v1) throws Exception {

                return v1.length();
            }

        });

        c.subtract(d).foreach(new VoidFunction<Tuple2<Integer, String>>() {
            public void call(Tuple2<Integer, String> t) throws Exception {
                // TODO Auto-generated method stub
                System.out.println("(" + t._1 + ", " + t._2 + ")");
            }
        });
    }

    /**
     * 取出RDD的前n个元素，以数组的形式返回
     */
    public static void demo4() {
        JavaSparkContext ctx = createContext();
        JavaRDD<String> a = ctx.parallelize(new ArrayList<String>(Arrays.asList("1", "4", "2", "3")));

        List<String> b = a.take(3);

        for (String c : b) {
            System.out.println(c);
        }

    }

    /**
     * 获得前几个最大值 output - hello 3
     */
    public static void demo5() {
        JavaSparkContext ctx = createContext();
        JavaRDD<String> a = ctx.parallelize(new ArrayList<String>(Arrays.asList("1", "hello", "2", "3")));
        List<String> b = a.top(2);
        for (String c : b) {
            System.out.println(c);
        }
    }

}