SparkRDD之cogroup和groupWith

cogroup和groupWith是一组非常强大的功能，允许使用键将最多3个键值RDD组合在一起。

示例如下：

package com.cb.spark.sparkrdd;

import java.util.Arrays;

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 scala.Tuple2;

public class CogroupExample {
	public static void main(String[] args) {
		SparkConf conf = new SparkConf().setAppName("Cogroup").setMaster("local");
		JavaSparkContext jsc = new JavaSparkContext(conf);

		JavaRDD<Integer> javaRDD = jsc.parallelize(Arrays.asList(1, 2, 1, 3), 2);

		JavaPairRDD<Integer, String> b = javaRDD.mapToPair(x -> new Tuple2<>(x, "b"));
		JavaPairRDD<Integer, String> c = javaRDD.mapToPair(x -> new Tuple2<>(x, "c"));

		// (1,([b, b],[c, c])) (3,([b],[c])) (2,([b],[c]))
		b.cogroup(c).foreach(x -> System.out.print(x + " "));

		JavaPairRDD<Integer, String> x = jsc.parallelizePairs(
				Arrays.asList(new Tuple2<Integer, String>(1, "apple"), new Tuple2<Integer, String>(2, "banana"),
						new Tuple2<Integer, String>(3, "orange"), new Tuple2<Integer, String>(4, "kiwi")),
				2);
		JavaPairRDD<Integer, String> y = jsc.parallelizePairs(
				Arrays.asList(new Tuple2<Integer, String>(5, "computer"), new Tuple2<Integer, String>(1, "laptop"),
						new Tuple2<Integer, String>(1, "desktop"), new Tuple2<Integer, String>(4, "ipad")),
				2);
		// (1,([apple],[laptop, desktop]))(3,([orange],[]))(5,([],[computer]))
		// (4,([kiwi],[ipad]))(2,([banana],[]))
		x.cogroup(y).foreach(m -> System.out.print(m));

		// (4,([kiwi],[ipad]))(2,([banana],[]))
		// (1,([apple],[laptop, desktop]))(3,([orange],[]))(5,([],[computer]))
		x.cogroup(y, b).foreach(m -> System.out.print(m));

		// (1,([apple],[laptop, desktop]))(3,([orange],[]))(5,([],[computer]))
		// (4,([kiwi],[ipad]))(2,([banana],[]))
		x.groupWith(y).foreach(m -> System.out.print(m));
		jsc.stop();
	}
}