Hadoop WordCount进阶

看了到Hadoop的代码，还是不知道他的执行流程，怎么办呢。我想到了日志，在hadoop的目录下，有log4j,那就用户Log4j来记录Hadoop的执行过程吧.

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;
import org.apache.log4j.Logger;

public class WordCount {

  public static Logger loger = Wloger.loger;

  /**
   * TokenizerMapper 继续自 Mapper<Object, Text, Text, IntWritable>
   *
   * [一个文件就一个map,两个文件就会有两个map]
   * map[这里读入输入文件内容 以" \t\n\r\f" 进行分割，然后设置 word ==> one 的key/value对]
   *
   * @param Object  Input key Type:
   * @param Text    Input value Type:
   * @param Text    Output key Type:
   * @param IntWritable Output value Type:
   *
   * Writable的主要特点是它使得Hadoop框架知道对一个Writable类型的对象怎样进行serialize以及deserialize.
   * WritableComparable在Writable的基础上增加了compareT接口，使得Hadoop框架知道怎样对WritableComparable类型的对象进行排序。
   *
   * @author yangchunlong.tw
   *
   */
  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 {
      loger.info("Map <key>"+key+"</key>");
      loger.info("Map <value>"+value+"</key>");
      StringTokenizer itr = new StringTokenizer(value.toString());
      while (itr.hasMoreTokens()) {
    	String wordstr = itr.nextToken();
        word.set(wordstr);
        loger.info("Map <word>"+wordstr+"</word>");
        context.write(word, one);
      }
    }
  }

  /**
   * IntSumReducer 继承自 Reducer<Text,IntWritable,Text,IntWritable>
   *
   * [不管几个Map,都只有一个Reduce,这是一个汇总]
   * reduce[循环所有的map值,把word ==> one 的key/value对进行汇总]
   *
   * 这里的key为Mapper设置的word[每一个key/value都会有一次reduce]
   *
   * 当循环结束后，最后的确context就是最后的结果.
   *
   * @author yangchunlong.tw
   *
   */
  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 {
      loger.info("Reduce <key>"+key+"</key>");
      loger.info("Reduce <value>"+values+"</key>");
      int sum = 0;
      for (IntWritable val : values) {
        sum += val.get();
      }
      result.set(sum);
      loger.info("Reduce <sum>"+sum+"</sum>");
      context.write(key, result);
    }
  }

  public static void main(String[] args) throws Exception {
    Configuration conf = new Configuration();
    String[] otherArgs = new GenericOptionsParser(conf, args).getRemainingArgs();
    /**
     * 这里必须有输入/输出
     */
    if (otherArgs.length != 2) {
      System.err.println("Usage: wordcount <in> <out>");
      System.exit(2);
    }
    Job job = new Job(conf, "word count");
    job.setJarByClass(WordCount.class);//主类
    job.setMapperClass(TokenizerMapper.class);//mapper
    job.setCombinerClass(IntSumReducer.class);//作业合成类
    job.setReducerClass(IntSumReducer.class);//reducer
    job.setOutputKeyClass(Text.class);//设置作业输出数据的关键类
    job.setOutputValueClass(IntWritable.class);//设置作业输出值类
    FileInputFormat.addInputPath(job, new Path(otherArgs[0]));//文件输入
    FileOutputFormat.setOutputPath(job, new Path(otherArgs[1]));//文件输出
    System.exit(job.waitForCompletion(true) ? 0 : 1);//等待完成退出.
  }
}

这里输出了每一次Map，每一次Reduce.结果如下:

f1 ==>Map Result
Map <key>0</key>
Map <value>ycl ycl is ycl good</key>
Map <word>ycl</word>
Map <word>ycl</word>
Map <word>is</word>
Map <word>ycl</word>
Map <word>good</word>

f1 ==>Reduce Result
Reduce <key>good</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1dfc547</key>
Reduce <sum>1</sum>
Reduce <key>is</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1dfc547</key>
Reduce <sum>1</sum>
Reduce <key>ycl</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1dfc547</key>
Reduce <sum>3</sum>

f2 ==>Map Result
Map <key>0</key>
Map <value>hello ycl hello lg</key>
Map <word>hello</word>
Map <word>ycl</word>
Map <word>hello</word>
Map <word>lg</word>

f2 ==>Reduce Result
Reduce <key>hello</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@10f6d3</key>
Reduce <sum>2</sum>
Reduce <key>lg</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@10f6d3</key>
Reduce <sum>1</sum>
Reduce <key>ycl</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@10f6d3</key>
Reduce <sum>1</sum>

f1,f2 ==> Reduce Result
Reduce <key>good</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1989f84</key>
Reduce <sum>1</sum>
Reduce <key>hello</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1989f84</key>
Reduce <sum>2</sum>
Reduce <key>is</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1989f84</key>
Reduce <sum>1</sum>
Reduce <key>lg</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1989f84</key>
Reduce <sum>1</sum>
Reduce <key>ycl</key>
Reduce <value>org.apache.hadoop.mapreduce.ReduceContext$ValueIterable@1989f84</key>
Reduce <sum>4</sum>

正常人应该能分析出map/reduce的执行机制，比如有两个输入文件，map/reduce是一个文件一个文件进行处理的，每map一个输入文件就会reduce一次，最后再进行总的reduce.
[注意这里不是一行处理一次，而是一个文件处理一次，没有进行split]