mapreduce编程

wordcount

测试：文件1 4.3m，文件100 430m；不设置combiner，设置combiner

单机环境

1.java

eclipse，maven一直没有成功，工程手动添加依赖

hadoop-common-2.7.3.jar
hadoop-mapreduce-client-core-2.7.3.jar
commons-cli-1.2.jar
hadoop-client-2.7.3.jar
hadoop-hdfs-2.7.3.jar

linux下编译class，打包jar

保存hadoop classpath，后续引用方便

tmp=`bin/hadoop classpath`
javac -classpath $tmp XMWordCount.java
MANIFEST.MF
Main-Class: com.XMWordCount
jar cfm XMWordCount1.jar com/MANIFEST.MF com/*.class
增加combiner，生成XMWordCount2.jar
jar cfm XMWordCount2.jar com/MANIFEST.MF com/*.class
bin/hadoop jar /home/xiumu/XMWordCount1.jar input1 output9

//XMWordCount.java
package com;

import java.io.IOException;
import java.util.StringTokenizer;

import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.IntWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.Mapper;
import org.apache.hadoop.mapreduce.Reducer;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.util.GenericOptionsParser;

public class XMWordCount {

  public static class TokenizerMapper 
       extends Mapper<Object, Text, Text, IntWritable>{
    
    private final static IntWritable one = new IntWritable(1);
    private Text word = new Text();
      
    public void map(Object key, Text value, Context context
                    ) throws IOException, InterruptedException {
      StringTokenizer itr = new StringTokenizer(value.toString());
      while (itr.hasMoreTokens()) {
        String s1 = itr.nextToken();
        StringBuilder s2 = new StringBuilder("");
        int len  = s1.length();
        for (int i = 0; i < len; ++i) {
          char c = s1.charAt(i);
          if (c >= 'a' && c <= 'z') s2.append(c);
          if (c >= 'A' && c <= 'Z') s2.append(c);
          if (c == '\'' || c == '-') s2.append(c);
        }
        if(!s2.toString().equals("")) {
        	word.set(s2.toString().toLowerCase());
            context.write(word, one);
        }
      }
    }
  }
  
  public static class IntSumReducer 
       extends Reducer<Text,IntWritable,Text,IntWritable> {
    private IntWritable result = new IntWritable();

    public void reduce(Text key, Iterable<IntWritable> values, 
                       Context context
                       ) throws IOException, InterruptedException {
      int sum = 0;
      for (IntWritable val : values) {
        sum += val.get();
      }
      result.set(sum);
      context.write(key, result);
    }
  }

  public static void main(String[] args) throws Exception {
    Configuration conf = new Configuration();
    String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
    if (otherArgs.length < 2) {
      System.err.println("Usage: XMwordcount <in> [<in>...] <out>");
      System.exit(2);
    }
    Job job = Job.getInstance(conf, "xiumu word count");
//    Job job = new Job();
    job.setJarByClass(XMWordCount.class);
    job.setMapperClass(TokenizerMapper.class);
//    job.setCombinerClass(IntSumReducer.class);
    job.setReducerClass(IntSumReducer.class);
    job.setOutputKeyClass(Text.class);
    job.setOutputValueClass(IntWritable.class);
    for (int i = 0; i < otherArgs.length - 1; ++i) {
      FileInputFormat.addInputPath(job, new Path(otherArgs[i]));
    }
    FileOutputFormat.setOutputPath(job,
      new Path(otherArgs[otherArgs.length - 1]));
    System.exit(job.waitForCompletion(true) ? 0 : 1);
  }
}

2.streaming，python3

bin/hadoop jar share/hadoop/tools/lib/hadoop-streaming-2.7.3.jar -input input1 -output output1 -mapper "/home/xiumu/mapper.py" -reducer "/home/xiumu/reduce.py"
bin/hadoop jar share/hadoop/tools/lib/hadoop-streaming-2.7.3.jar -input input1 -output output1 -mapper "/home/xiumu/mapper.py" -reducer "/home/xiumu/reduce.py" -combiner "/home/xiumu/reduce.py"

mapper.py

#!/usr/bin/python3
import sys
for line in sys.stdin :
  words = line.strip().split()
  for word in words :
    newword = str()
    for c in word : 
      if c >= 'a' and c <= 'z' :
        newword += c            
      if c >= 'A' and c <= 'Z' :
        newword += c            
      if c == '\'' or c == '-' :
        newword += c            
    newword = newword.strip().lower()
    if newword == '' :
      continue          
    print("%s %d" % (newword, 1))

reducer.py

#!/usr/bin/python3
import sys
(last_key, count) = (None, 0)
for line in sys.stdin :
  (key, value) = line.strip().split()
  if last_key == key : 
    count += int(value)
  else :
    if last_key != None :
      print("%s %d" % (last_key, count))
    (last_key, count) = (key, int(value)) 
if last_key != None :
  print("%s %d" % (last_key, count))

结论：

java比streaming快好多，streaming可以选自己喜欢的脚本语言，简单，但是效率低，而且对整个过程的控制不如java

combiner，比较大的影响效率，尤其是在reduce阶段会快很多，原因很显然，不表

而且java运行时，机器负载较低