mapreduce 二次排序

目标： 

 

输入数据：

1. sort1   1  
2. sort2   3  
3. sort2   88  
4. sort2   54  
5. sort1   2  
6. sort6   22  
7. sort6   888  
8. sort6   58

 

 

 

       输出数据：

 

1. sort1   1,2  
2. sort2   3,54,88  
3. sort6   22,58,888  

既然是二次排序，所以上述的输入数据必须在同一分区中，不然达不到我们想要的二次排序效果。MapReduce框架默认是对key进行排序的，这里我们需要自定义排序方法对key进行排序。

 

先理解MapReduce处理数据的流程：

（这是在另一个大神的博客看到的     http://blog.csdn.net/lzm1340458776/article/details/42875751）

 

我们应该先深刻的理解MapReduce处理数据的整个流程，这是最基础的，不然的话是不可能找到解决问题的思路的。我描述一下MapReduce处理数据的大概流程：首先，MapReduce框架通过getSplits()方法实现对原始文件的切片之后，每一个切片对应着一个MapTask，InputSplit输入到map()函数进行处理，中间结果经过环形缓冲区的排序，然后分区、自定义二次排序(如果有的话)和合并，再通过Shuffle操作将数据传输到reduce Task端，reduce端也存在着缓冲区，数据也会在缓冲区和磁盘中进行合并排序等操作，然后对数据按照key值进行分组，然后每处理完一个分组之后就会去调用一次reduce()函数，最终输出结果。大概流程 我画了一下，如下图：

大概解决思路：

map端排序：

首先将输入数据放到同一分区中并对其进行排序，因为MapReduce排序只对key进行排序，所以我们需要一个组合key，以[key,value]的形式输入，map输入的数据类型结构大概为{[key,value],value}:

 

1. {[sort1,1],1}  
2. {[sort2,3],3}  
3. {[sort2,88],88}  
4. {[sort2,54],54}  
5. {[sort1,2],2}  
6. {[sort6,22],22}  
7. {[sort6,888],888}  
8. {[sort6,58],58}  

温馨提示：请务必保证map端排序结果正确，不然不管怎么分组都得不到你想要的结果（本人就踩过这坑，费了好长时间才找到原因）。

 

map端排序后的结果应该如下：

 

1. sort1:1,
   sort1:2,
   sort2:3,
   sort2:54,
   sort2:88,
   sort6:22,
   sort6:58,
   sort6:888,
2.   

reduce端分组：

 

前面提到过，reduce方法执行的次数是由分组数决定的。在reduce端按组合key的第一个字段进行分组，然后再对对应的value值进行相应的处理则可以达到我们想要的结果。

 

1. sort1   1,2  
2. sort2   3,54,88  
3. sort6   22,58,888  

 

 

 

 

具体的代码实现如下：

自定义排序类：

 

package com.secondSortJob;

import com.sun.deploy.util.ArrayUtil;
import org.apache.commons.lang.ArrayUtils;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.io.WritableComparable;

import java.io.DataInput;
import java.io.DataOutput;
import java.io.IOException;


/**
 * Created by Administrator on 2018/1/28 0028.
 */
public class SecondKeySort implements WritableComparable {
    private Text key = new Text();
    private IntWritable count = new IntWritable();

    public SecondKeySort(Text key, IntWritable count) {
        this.key = key;
        this.count = count;
    }

    public SecondKeySort() {
    }

    public Text getKey() {

        return key;
    }

    public void setKey(Text key) {
        this.key = key;
    }

    public IntWritable getCount() {
        return count;
    }

    public void setCount(IntWritable count) {
        this.count = count;
    }

    public void write(DataOutput dataOutput) throws IOException {
//        new Text("["+this.key.toString()+","+this.count.toString()+"]");
        this.key.write(dataOutput);
        this.count.write(dataOutput);
    }

    public void readFields(DataInput dataInput) throws IOException {
        this.key.readFields(dataInput);
        this.count.readFields(dataInput);
    }

    public int compareTo(SecondKeySort o) {
        String keyStr1 = this.key.toString().substring(0,this.key.toString().length());
        String keyStr2 = o.getKey().toString().substring(0,o.getKey().toString().length());

        String[] keyStr1Arr = keyStr1.split("\t");
        String[] keyStr2Arr = keyStr2.split("\t");

        String keyStr1Text = keyStr1Arr [0];
        String keyStr2Text = keyStr2Arr [0];

        if(!keyStr1Text.equals(keyStr2Text)){
            System.out.println(keyStr1Text.compareTo(keyStr2Text));
            return keyStr1Text.compareTo(keyStr2Text) ;
        }else{
            return  Integer.parseInt(this.count.toString()) - Integer.parseInt(o.getCount().toString()) ;
        }
    }

    @Override
    public String toString() {
        return  this.key.toString() +"\t" +this.count.toString();
    }

}

mapper任务类：

 

 

package com.secondSortJob;

import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;

import java.io.IOException;

/**
 * Created by Administrator on 2018/1/28 0028.
 */
public class SecondSortMapper extends Mapper {
    private SecondKeySort secondKeySort = new SecondKeySort();
    @Override
    protected void map(LongWritable key, Text value, Context context) throws IOException, InterruptedException {
        String[] arrs = value.toString().split("\t");
        secondKeySort.setKey(new Text(arrs[0]));
        secondKeySort.setCount(new IntWritable(Integer.parseInt(arrs[1].trim())));
        context.write(secondKeySort,new IntWritable(Integer.parseInt(arrs[1].trim())));
        System.out.println(secondKeySort.toString()+"  "+arrs[1]);
    }
}

 

reduce任务类：

package com.secondSortJob;

import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;

import java.io.IOException;


/**
 * Created by Administrator on 2018/1/28 0028.
 */
public class SecondSortReducer extends Reducer {
    @Override
    protected void reduce(SecondKeySort key, Iterable values, Context context) throws IOException, InterruptedException {
        System.out.println("reduce action");
        String str = key.getKey().toString() +":";
        while (values.iterator().hasNext()){

            str = str + values.iterator().next()+",";
        }
        context.write(new Text(str),NullWritable.get());
    }
}

自定义分区类：（保证输入数据在同一个分区，分区数和Reduce任务数相等）

 

 

package com.secondSortJob;

import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.mapred.JobConf;
import org.apache.hadoop.mapreduce.Partitioner;


/**
 * Created by Administrator on 2018/1/29 0029.
 */
public class GroupingPartitioner extends Partitioner {
    public int getPartition(SecondKeySort secondKeySort, IntWritable intWritable, int i) {
        return secondKeySort.getKey().toString().hashCode()%i ;
    }
}

 

分组函数类：（可以实现RawComparator接口，也可以继承WritableComparator类）

 

package com.secondSortJob;

import org.apache.hadoop.io.RawComparator;
import org.apache.hadoop.io.WritableComparator;

/**
 * Created by Administrator on 2018/1/28 0028.
 */
public class GroupingComparator implements RawComparator {
    public int compare(byte[] bytes, int i, int i1, byte[] bytes1, int i2, int i3) {
        int compareBytes = WritableComparator.compareBytes(bytes, i, 8, bytes1, i2, 8);
        return compareBytes;
    }
    public int compare(SecondKeySort o1, SecondKeySort o2) {
        return o1.getKey().compareTo(o2.getKey());
    }
}

 

 

 

 

 

主程序加载类：

 

package com.secondSortJob;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.NullWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;

import java.io.IOException;
import java.net.URI;

/**
 * 二次排序
 * Created by Administrator on 2018/1/28 0028.
 */
public class SecondSortJob {
    private  static     Configuration configuration = new Configuration();
    private  static     String master = "hdfs://192.168.8.222";
    private  static      String masterFS = "hdfs://192.168.8.222:9000";
    private    static      FileSystem fs = null ;

    public static void main(String[] args) throws IOException, ClassNotFoundException, InterruptedException {
        configuration.set("yarn.resourcemanager.hostname",master);
        fs =  FileSystem.get(URI.create(masterFS),configuration);
        Job job = Job.getInstance(configuration);

        String inputPath = "/lsw/secondSort.txt";
        String secondSortResult = "/lsw/secondSortResult" ;


        if(fs.isDirectory(new Path(masterFS+secondSortResult)))
            fs.delete(new Path(masterFS+secondSortResult));

        job.setMapOutputKeyClass(SecondKeySort.class);
        job.setMapOutputValueClass(IntWritable.class);

        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(NullWritable.class);

        job.setJarByClass(SecondSortJob.class);
        job.setMapperClass(SecondSortMapper.class);
        job.setReducerClass(SecondSortReducer.class);

        job.setPartitionerClass(GroupingPartitioner.class);
        job.setGroupingComparatorClass(GroupingComparator.class);
//        job.setNumReduceTasks(3);
        FileInputFormat.addInputPath(job,new Path(masterFS+inputPath));
        FileOutputFormat.setOutputPath(job,new Path(masterFS+secondSortResult));

        job.waitForCompletion(true);
    }
}