当前位置: 代码迷 >> 综合 >> Mapper reduce
  详细解决方案

Mapper reduce

热度:39   发布时间:2023-10-26 20:36:16.0

MapReduce对应的java类:

package com.paic;import java.io.IOException;  
import java.util.Iterator;  
import java.util.StringTokenizer;  
import org.apache.hadoop.fs.Path;  
import org.apache.hadoop.io.IntWritable;  
import org.apache.hadoop.io.LongWritable;  
import org.apache.hadoop.io.Text;  
import org.apache.hadoop.mapred.FileInputFormat;  
import org.apache.hadoop.mapred.FileOutputFormat;  
import org.apache.hadoop.mapred.JobClient;  
import org.apache.hadoop.mapred.JobConf;  
import org.apache.hadoop.mapred.MapReduceBase;  
import org.apache.hadoop.mapred.Mapper;  
import org.apache.hadoop.mapred.OutputCollector;  
import org.apache.hadoop.mapred.Reducer;  
import org.apache.hadoop.mapred.Reporter;  
import org.apache.hadoop.mapred.TextInputFormat;  
import org.apache.hadoop.mapred.TextOutputFormat;  
/** *  * 描述:WordCount explains by Felix * @author Hadoop Dev Group */  
public class WordCount  
{  /** * MapReduceBase类:实现了Mapper和Reducer接口的基类(其中的方法只是实现接口,而未作任何事情) * Mapper接口: * WritableComparable接口:实现WritableComparable的类可以相互比较。所有被用作key的类应该实现此接口。 * Reporter 则可用于报告整个应用的运行进度,本例中未使用。  *  */  public static class Map extends MapReduceBase implements  Mapper<LongWritable, Text, Text, IntWritable>  {  /** * LongWritable, IntWritable, Text 均是 Hadoop 中实现的用于封装 Java 数据类型的类,这些类实现了WritableComparable接口, * 都能够被串行化从而便于在分布式环境中进行数据交换,你可以将它们分别视为long,int,String 的替代品。 */  private final static IntWritable one = new IntWritable(10);  private Text word = new Text();  /** * Mapper接口中的map方法: * void map(K1 key, V1 value, OutputCollector<K2,V2> output, Reporter reporter) * 映射一个单个的输入k/v对到一个中间的k/v对 * 输出对不需要和输入对是相同的类型,输入对可以映射到0个或多个输出对。 * OutputCollector接口:收集Mapper和Reducer输出的<k,v>对。 * OutputCollector接口的collect(k, v)方法:增加一个(k,v)对到output */  public void map(LongWritable key, Text value,  OutputCollector<Text, IntWritable> output, Reporter reporter)  throws IOException  {  //String line = value.toString();  StringTokenizer tokenizer = new StringTokenizer(line);  while (tokenizer.hasMoreTokens())  {  word.set(tokenizer.nextToken());  output.collect(word, one);  }  }  }  public static class Reduce extends MapReduceBase implements  Reducer<Text, IntWritable, Text, IntWritable>  {  public void reduce(Text key, Iterator<IntWritable> values,  OutputCollector<Text, IntWritable> output, Reporter reporter)  throws IOException  {  int sum = 0;  while (values.hasNext())  {  sum += values.next().get();  }  output.collect(key, new IntWritable(sum));  }  }  public static void main(String[] args) throws Exception  {  /** * JobConf:map/reduce的job配置类,向hadoop框架描述map-reduce执行的工作 * 构造方法:JobConf()、JobConf(Class exampleClass)、JobConf(Configuration conf)等 */  
//     System.setProperty("HADOOP_HOME", "D:\\HADOOP_HOME");
//     System.setProperty("path", "D:\\HADOOP_HOME\\bin");JobConf conf = new JobConf(WordCount.class);  conf.setJobName("wordcount");           //设置一个用户定义的job名称  conf.setOutputKeyClass(Text.class);    //为job的输出数据设置Key类  conf.setOutputValueClass(IntWritable.class);   //为job输出设置value类  conf.setMapperClass(Map.class);         //为job设置Mapper类  conf.setCombinerClass(Reduce.class);      //为job设置Combiner类  conf.setReducerClass(Reduce.class);        //为job设置Reduce类  conf.setInputFormat(TextInputFormat.class);    //为map-reduce任务设置InputFormat实现类  conf.setOutputFormat(TextOutputFormat.class);  //为map-reduce任务设置OutputFormat实现类  /** * InputFormat描述map-reduce中对job的输入定义 * setInputPaths():为map-reduce job设置路径数组作为输入列表 * setInputPath():为map-reduce job设置路径数组作为输出列表 */  FileInputFormat.setInputPaths(conf, new Path("D:\\out3.txt"));  FileOutputFormat.setOutputPath(conf, new Path("D:\\out4"));  JobClient.runJob(conf);         //运行一个job  }  
}  

 

  相关解决方案