map/reduce實現 排序

林六天發表於2014-07-07
原文網址 : https://www.cnblogs.com/6tian/p/3829169.html
排序
  1 import java.io.IOException;
  2 
  3 import org.apache.hadoop.conf.Configuration;
  4 import org.apache.hadoop.conf.Configured;
  5 import org.apache.hadoop.fs.Path;
  6 import org.apache.hadoop.io.IntWritable;
  7 import org.apache.hadoop.io.LongWritable;
  8 import org.apache.hadoop.io.Text;
  9 import org.apache.hadoop.mapreduce.Job;
 10 import org.apache.hadoop.mapreduce.Mapper;
 11 import org.apache.hadoop.mapreduce.Partitioner;
 12 import org.apache.hadoop.mapreduce.Reducer;
 13 import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
 14 import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
 15 import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
 16 import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat;
 17 import org.apache.hadoop.util.Tool;
 18 import org.apache.hadoop.util.ToolRunner;
 19 public class Sort extends Configured implements Tool {
 20     /*
 21      * 排序
 22      * 輸入格式：每個資料佔一行
 23      * 輸出格式：
 24      * 1 21
 25      * 2 32
 26      * 3 62
 27      * 設計思路：
 28      * 使用reduce自帶的預設排序規則。MapReduce按照key值進行排序。如果Key值為Intwritable型別，則按照數字大小排序
 29      * 如果key值為Text型別，則按照字典順序對字串進行排序。
 30      * 注意：要重寫Partition函式。Reduce排序只能保證自己區域性的資料順序，並不能保證全域性的。
 31      * */
 32     public static class Map extends Mapper<LongWritable,Text,IntWritable,IntWritable>{
 33         private IntWritable line=new IntWritable();
 34         public void map(LongWritable key,Text value,Context context)throws IOException,InterruptedException{
 35             line.set(Integer.parseInt(value.toString()));
 36             context.write(line, new IntWritable(1));            
 37         }
 38         
 39     }
 40     
 41     public static class Reduce extends Reducer<IntWritable,IntWritable,IntWritable,IntWritable>{
 42         private IntWritable num=new IntWritable(1);
 43         public void reduce(IntWritable key,Iterable<IntWritable> values,Context context)throws IOException,InterruptedException{
 44             for(IntWritable var:values){
 45             context.write(num, key);
 46             num=new IntWritable(num.get()+1);
 47             }
 48         }
 49         
 50     }
 51     
 52     public static class Partition extends Partitioner<IntWritable ,IntWritable>{
 53 
 54         @Override
 55         public int getPartition(IntWritable key, IntWritable value, int numPartitions) {
 56             // TODO Auto-generated method stub
 57             System.out.println(numPartitions);
 58             int maxnum=65223;
 59             int bound=maxnum/numPartitions+1;
 60             for(int i=0;i<numPartitions;i++)
 61             {
 62                 if(key.get()>=bound*(i-1)&&key.get()<=bound*i)
 63                 {
 64                     return i;
 65                 }
 66             }
 67             return 0;
 68         }
 69         
 70     }
 71     
 72     public int run(String[] args)throws Exception{
 73         Configuration conf=new Configuration();
 74         Job job=new Job(conf,"Sort");
 75         job.setJarByClass(Sort.class);
 76         
 77         job.setOutputKeyClass(IntWritable.class);
 78         job.setOutputValueClass(IntWritable.class);
 79         
 80         
 81         job.setMapperClass(Map.class);
 82         job.setReducerClass(Reduce.class);
 83         job.setPartitionerClass(Partition.class);
 84         
 85         job.setInputFormatClass(TextInputFormat.class);
 86         job.setOutputFormatClass(TextOutputFormat.class);
 87         
 88         FileInputFormat.addInputPath(job, new Path(args[0]));
 89         FileOutputFormat.setOutputPath(job, new Path(args[1]));
 90         
 91         boolean success=job.waitForCompletion(true);
 92         return success?0:1;
 93     }
 94     
 95     public static void main(String[] args)throws Exception{
 96         int ret=ToolRunner.run(new Sort(), args);
 97         System.exit(ret);
 98     }
 99 
100 }

 

相關文章