Spark Streaming(2)：Join

package com.pyrrha.examples;

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

import org.apache.spark.SparkConf;
import org.apache.spark.api.java.JavaPairRDD;
import org.apache.spark.api.java.JavaSparkContext;
import org.apache.spark.api.java.Optional;

import scala.Tuple2;

public class JoinTransformation {
	
	public static void main(String[] args) {
		SparkConf conf = new SparkConf().setAppName("TransformationTest").setMaster("local");
		JavaSparkContext sc = new JavaSparkContext(conf);
		
		join(sc);
		
		sc.close();
	}
	/**
	 * @return：
	    ListA:
		[(name,Jos), (name,Marry), (name1,Tom)]
	    ListB
		[(name,Jos), (name,Black), (name2,David)]
		
		-----------------join-------------------
		[(name,(Jos,Jos)), (name,(Jos,Black)), (name,(Marry,Jos)), (name,(Marry,Black))]
		
		-----------------leftOuterJoin-------------------
		[(name,(Jos,Optional[Jos])), (name,(Jos,Optional[Black])), (name,(Marry,Optional[Jos])), (name,(Marry,Optional[Black])), (name1,(Tom,Optional.empty))]
		
		-----------------rightOuterJoin-------------------
		[(name,(Optional[Jos],Jos)), (name,(Optional[Jos],Black)), (name,(Optional[Marry],Jos)), (name,(Optional[Marry],Black)), (name2,(Optional.empty,David))]
		
		-----------------fullOuterJoin-------------------
		[(name,(Optional[Jos],Optional[Jos])), (name,(Optional[Jos],Optional[Black])), (name,(Optional[Marry],Optional[Jos])), (name,(Optional[Marry],Optional[Black])), (name2,(Optional.empty,Optional[David])), (name1,(Optional[Tom],Optional.empty))]
	 * @param sc
	 */
	public static void join(JavaSparkContext sc) {
		List<Tuple2<String, String>> listA = new ArrayList<Tuple2<String, String>>();
		listA.add(new Tuple2<String, String>("name","Jos"));
		listA.add(new Tuple2<String, String>("name","Marry"));
		listA.add(new Tuple2<String, String>("name1","Tom"));
		List<Tuple2<String, String>> listB = new ArrayList<Tuple2<String, String>>();
		listB.add(new Tuple2<String, String>("name","Jos"));
		listB.add(new Tuple2<String, String>("name","Black"));
		listB.add(new Tuple2<String, String>("name2","David"));
		
		JavaPairRDD<String, String> RDDA = sc.parallelizePairs(listA);
		JavaPairRDD<String, String> RDDB = sc.parallelizePairs(listB);
		System.out.println("-----------------data-------------------");
		System.out.println("ListA:");
		System.out.println(RDDA.collect());
		System.out.println("\n");
		System.out.println("ListB");
		System.out.println(RDDB.collect());
		System.out.println("\n");
		
		
		JavaPairRDD<String, Tuple2<String, String>> RDDC = RDDA.join(RDDB);
		System.out.println("-----------------join-------------------");
		System.out.println(RDDC.collect());
		System.out.println("\n");
		
		JavaPairRDD<String, Tuple2<String, Optional<String>>> RDDD = RDDA.leftOuterJoin(RDDB);
		System.out.println("-----------------leftOuterJoin-------------------");
		System.out.println(RDDD.collect());
		System.out.println("\n");
		
		JavaPairRDD<String, Tuple2<Optional<String>, String>> RDDE = RDDA.rightOuterJoin(RDDB);
		System.out.println("-----------------rightOuterJoin-------------------");
		System.out.println(RDDE.collect());
		System.out.println("\n");
		
		JavaPairRDD<String, Tuple2<Optional<String>, Optional<String>>> RDDF = RDDA.fullOuterJoin(RDDB);
		System.out.println("-----------------fullOuterJoin-------------------");
		System.out.println(RDDF.collect());
		System.out.println("\n");
		
	}
	
	
	

}

 

转载于:https://my.oschina.net/jos/blog/1504646