MapReduce综合案例(自定义类+分区+优化器+分组)

数据：

班级|姓名|学科|成绩（存在重复）

ClassA|Jack|Math|95
ClassA|Jack|Math|95
ClassA|Jack|English|95
ClassA|Jack|English|95
ClassA|Tom|Math|85
ClassA|Tom|Math|85
ClassA|Tom|English|85
ClassA|Tom|English|85
ClassA|Jerry|Math|80
ClassA|Jerry|Math|80
ClassA|Jerry|English|80
ClassA|Jerry|English|80
ClassA|Bob|Math|60
ClassA|Bob|Math|60
ClassA|Bob|English|60
ClassA|Bob|English|60
ClassA|Alice|Math|90
ClassA|Alice|Math|90
ClassA|Alice|English|90
ClassA|Alice|English|90
ClassA|Rose|Math|90
ClassA|Rose|Math|90
ClassA|Rose|English|90
ClassA|Rose|English|90
ClassB|Jack|Math|90
ClassB|Jack|Math|90
ClassB|Jack|English|95
ClassB|Jack|English|95
ClassB|Tom|Math|85
ClassB|Tom|Math|85
ClassB|Tom|English|85
ClassB|Tom|English|85
ClassB|Jerry|Math|80
ClassB|Jerry|Math|80
ClassB|Jerry|English|80
ClassB|Jerry|English|80
ClassB|Bob|Math|60
ClassB|Bob|Math|60
ClassB|Bob|English|60
ClassB|Bob|English|60
ClassB|Alice|Math|90
ClassB|Alice|Math|90
ClassB|Alice|English|90
ClassB|Alice|English|90
ClassB|Rose|Math|99
ClassB|Rose|Math|99
ClassB|Rose|English|98
ClassB|Rose|English|98

需求：

将每个班级分别写出到不同的文件，内容是每个学科的前两名学生记录

处理结果1：

ClassB|Rose|English|98
ClassB|Jack|English|95
ClassB|Rose|Math|99
ClassB|Jack|Math|90

处理结果2：

ClassA|Jack|English|95
ClassA|Rose|English|90
ClassA|Jack|Math|95
ClassA|Alice|Math|90

分析：

1. 不同的班级写出到不同的文件需要自定义Partitioner来实现自定义分区
2. 数据存在重复为了减轻reducer端压力则需要自定义combiner优化器用来去重
3. 为了保证相同学科分到一组则需要自定义WritableCompatator用来分组（当然这里是为了用而用本身自定义类就已经实现）
4. 最后就是自定义类需要实现序列化和比较功能

代码实现：

package mr;

import org.apache.commons.beanutils.BeanUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.*;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Partitioner;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;
import java.lang.reflect.InvocationTargetException;
import java.util.StringTokenizer;

/**
 * @ClassName: TopNDriver
 * @Description: 
 * @Author: xuezhouyi
 * @Version: V1.0
 **/
public class TopNDriver {
	/* 学生类：必须实现序列化和比较功能 */
	public static class Student implements WritableComparable<Student> {
		private String stuClass;
		private String stuName;
		private String stuCourse;
		private Integer stuScore;

		public String getStuClass() {
			return stuClass;
		}

		public void setStuClass(String stuClass) {
			this.stuClass = stuClass;
		}

		public String getStuName() {
			return stuName;
		}

		public void setStuName(String stuName) {
			this.stuName = stuName;
		}

		public String getStuCourse() {
			return stuCourse;
		}

		public void setStuCourse(String stuCourse) {
			this.stuCourse = stuCourse;
		}

		public Integer getStuScore() {
			return stuScore;
		}

		public void setStuScore(Integer stuScore) {
			this.stuScore = stuScore;
		}

		@Override
		public String toString() {
			return this.stuClass + "|" + this.stuName + "|" + this.stuCourse + "|" + this.stuScore;
		}

		@Override
		public int compareTo(Student o) {
			/* 学科+分数比较，保证相同学科下分数高的降序 */
			int compare = this.getStuCourse().compareTo(o.getStuCourse());
			if (compare == 0)
				return -this.getStuScore().compareTo(o.getStuScore());
			return compare;
		}

		/* 序列化 */
		@Override
		public void write(DataOutput out) throws IOException {
			out.writeUTF(this.stuClass);
			out.writeUTF(this.stuName);
			out.writeUTF(this.stuCourse);
			out.writeInt(this.stuScore);
		}

		/* 反序列化 */
		@Override
		public void readFields(DataInput in) throws IOException {
			this.stuClass = in.readUTF();
			this.stuName = in.readUTF();
			this.stuCourse = in.readUTF();
			this.stuScore = in.readInt();
		}
	}

	public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {
		Configuration conf = new Configuration();
		Job job = Job.getInstance(conf);
		job.setJarByClass(TopNDriver.class);
		job.setMapperClass(MyMapper.class);
		job.setReducerClass(MyReducer.class);
		/* 设置map输出类型 */
		job.setMapOutputKeyClass(Student.class);
		job.setMapOutputValueClass(NullWritable.class);
		/* 设置reduce输出类型 */
		job.setOutputKeyClass(Student.class);
		job.setOutputValueClass(NullWritable.class);
		/* 设置分区：两个分区 */
		job.setPartitionerClass(MyPartitioner.class);
		job.setNumReduceTasks(2);
		/* 设置聚合 */
		job.setCombinerClass(MyCombiner.class);
		/* 设置分组 */
		job.setGroupingComparatorClass(MyGroup.class);
		/* 设置输入输出路径 */
		FileInputFormat.setInputPaths(job, new Path(args[0]));
		FileOutputFormat.setOutputPath(job, new Path(args[1]));
		/* 先清理再输出 */
		FileSystem fileSystem = FileSystem.get(conf);
		if (fileSystem.exists(new Path(args[1])))
			fileSystem.delete(new Path(args[1]), true);
		/* 退出 */
		System.exit(job.waitForCompletion(true) ? 0 : 1);
	}

	/* mapTask类 */
	private static class MyMapper extends Mapper<LongWritable, Text, Student, NullWritable> {
		Student k = new Student();

		@Override
		protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
			/* 数据拆分 */
			StringTokenizer tokenizer = new StringTokenizer(value.toString(), "|");

			/* 数据清洗 */
			if (tokenizer.countTokens() != 4)
				return;

			/* 数据封装 */
			String str = "";
			str = (String) tokenizer.nextElement();
			k.setStuClass(str);
			str = (String) tokenizer.nextElement();
			k.setStuName(str);
			str = (String) tokenizer.nextElement();
			k.setStuCourse(str);
			Integer score = Integer.parseInt((String) tokenizer.nextElement());
			k.setStuScore(score);

			/* 数据发送 */
			context.write(k, NullWritable.get());
		}
	}

	/* 分区类：按照班级进行分区 */
	private static class MyPartitioner extends Partitioner<Student, NullWritable> {
		@Override
		public int getPartition(Student student, NullWritable nullWritable, int numPartitions) {
			/* 根据班级的哈希值取余进行分区 */
			return (student.getStuClass().hashCode() & Integer.MAX_VALUE) % numPartitions;
		}
	}

	/* 局部聚合类：实现数据去重 */
	private static class MyCombiner extends Reducer<Student, NullWritable, Student, NullWritable> {
		Student k = new Student();

		@Override
		protected void reduce(Student key, Iterable<NullWritable> values, Context context) throws IOException, InterruptedException {
			/* 数据封装：需要注意类必须是public的 */
			try {
				BeanUtils.copyProperties(k, key);
			} catch (IllegalAccessException e) {
				e.printStackTrace();
			} catch (InvocationTargetException e) {
				e.printStackTrace();
			}

			/* 数据发送 */
			context.write(k, NullWritable.get());
		}
	}

	/* 分组类 */
	private static class MyGroup extends WritableComparator {
		public MyGroup() {
			super(Student.class, true);
		}

		@Override
		public int compare(WritableComparable a, WritableComparable b) {
			Student stuA = (Student) a;
			Student stuB = (Student) b;
			/* 相同学科分到一组 */
			return stuA.getStuCourse().compareTo(stuB.getStuCourse());
		}
	}

	/* reduceTask类 */
	private static class MyReducer extends Reducer<Student, NullWritable, Student, NullWritable> {
		@Override
		protected void reduce(Student key, Iterable<NullWritable> values, Context context) throws IOException, InterruptedException {
			int i = 0;
			for (NullWritable value : values) {
				/* 写出到HDFS */
				context.write(key, NullWritable.get());

				/* 每门课程的前两名 */
				if (i++ == 1)
					return;
			}
		}
	}
}