Spark WordCount的两种方式

Spark WordCount的两种方式。

语言：Java

工具：Idea

项目：Java Maven

pom.xml如下：

<properties>
        <spark.version>1.2.0</spark.version>
    </properties>

    <dependencies>
        <dependency>
            <groupId>org.apache.spark</groupId>
            <artifactId>spark-core_2.10</artifactId>
            <version>${spark.version}</version>
        </dependency>
    </dependencies>

第一种方式，比较常规的按部就班的

package pairs;

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.FlatMapFunction;
import org.apache.spark.api.java.function.Function2;
import org.apache.spark.api.java.function.PairFunction;
import org.apache.spark.api.java.function.VoidFunction;
import scala.Tuple2;
import java.util.Arrays;

public class WordCount1 {
    public static void main(String[] args) {
        SparkConf conf = new SparkConf().setMaster("local").setAppName("wordcount1");
        JavaSparkContext sc = new JavaSparkContext(conf);

        String filename = "D:\\tmp\\words.txt";
        JavaRDD<String> input = sc.textFile(filename);
        JavaRDD<String> lines = input.flatMap(new FlatMapFunction<String, String>() {
            public Iterable<String> call(String s) throws Exception {
                return Arrays.asList(s.split(" "));
            }
        });

        //pairs
        JavaPairRDD<String,Integer> pairs = lines.mapToPair(new PairFunction<String, String, Integer>() {
            public Tuple2<String, Integer> call(String s) throws Exception {
                return new Tuple2<String, Integer>(s,1);
            }
        });

        //reduce
        JavaPairRDD<String,Integer> counts = pairs.reduceByKey(new Function2<Integer, Integer, Integer>() {
            public Integer call(Integer x, Integer y) throws Exception {
                return x+y;
            }
        });

        //output
        counts.foreach(new VoidFunction<Tuple2<String, Integer>>() {
            public void call(Tuple2<String, Integer> tuple2) throws Exception {
                System.out.println(tuple2);
            }
        });

        sc.stop();

    }
}

代码输出：

(rose,2)
(jack,3)

第二种更为简洁

package pairs;

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

import java.util.Arrays;
import java.util.Map;

public class WordCount2 {

    public static void main(String[] args) {
        SparkConf conf = new SparkConf().setMaster("local").setAppName("wordcount2");
        JavaSparkContext sc = new JavaSparkContext(conf);

        String filename = "D:\\tmp\\words.txt";
        JavaRDD<String> input = sc.textFile(filename);
        JavaRDD<String> lines = input.flatMap(new FlatMapFunction<String, String>() {
            public Iterable<String> call(String s) throws Exception {
                return Arrays.asList(s.split(" "));
            }
        });

        Map<String,Long> result = lines.countByValue();
        System.out.println(result);

        sc.stop();
    }
}

代码输出：

{rose=2, jack=3}

通过对比可以发现，第一种方式一直都是转化操作，最后打印的是Tuple2；而第二种方式变成了行动操作，直接输出Map<String,Long>。

具体有什么区别，或者效率上有啥不同，待后续深入学习。

参考资料：

《Spark快速大数据分析》