Hadoop的MR作业支持链式处理，类似在一个生产牛奶的流水线上，每一个阶段都有特定的任务要处理，比如提供牛奶盒，装入牛奶，封盒，打印出厂日期，等等，通过这样进一步的分工，从而提高了生产效率，那么在我们的Hadoop的MapReduce中也是如此，支持链式的处理方式，这些Mapper像Linux管道一样，前一个Mapper的输出结果直接重定向到下一个Mapper的输入，形成一个流水线，而这一点与Lucene和Solr中的Filter机制是非常类似的，Hadoop项目源自Lucene，自然也借鉴了一些Lucene中的处理方式。
举个例子，比如处理文本中的一些禁用词，或者敏感词，等等，Hadoop里的链式操作，支持的形式类似正则Map+ Rrduce Map*，代表的意思是全局只能有一个唯一的Reduce，但是在Reduce的前后是可以存在无限多个Mapper来进行一些预处理或者善后工作的。
下面来看下的散仙今天的测试例子，先看下我们的数据，以及需求。
数据如下：
手机 5000 电脑 2000 衣服 300 鞋子 1200 裙子 434 手套 12 图书 12510 小商品 5 小商品 3 订餐 2
需求是：
/**
* 需求：
* 在第一个Mapper里面过滤大于10000万的数据
* 第二个Mapper里面过滤掉大于100-10000的数据
* Reduce里面进行分类汇总并输出
* Reduce后的Mapper里过滤掉商品名长度大于3的数据
*/
预计处理完的结果是：
手套 12
订餐 2
散仙的hadoop版本是1.2的,在1.2的版本里，hadoop支持新的API，但是链式的ChainMapper类和ChainReduce类却不支持新 的，新的在hadoop2.x里面可以使用，差别不大，散仙今天给出的是旧的API的，需要注意一下。
代码如下：
package com.qin.test.hadoop.chain; import java.io.IOException; import java.util.Iterator; import org.apache.hadoop.conf.Configuration; import org.apache.hadoop.fs.FileSystem; import org.apache.hadoop.fs.Path; import org.apache.hadoop.io.LongWritable; import org.apache.hadoop.io.Text; 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.lib.ChainMapper; import org.apache.hadoop.mapred.lib.ChainReducer; import org.apache.hadoop.mapreduce.lib.input.FileInputFormat; import org.apache.hadoop.mapreduce.lib.input.TextInputFormat; import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat; import org.apache.hadoop.mapreduce.lib.output.TextOutputFormat; import com.qin.reducejoin.NewReduceJoin2; /** * * 测试Hadoop里面的 * ChainMapper和ReduceMapper的使用 * * @author qindongliang * @date 2014年5月7日 * * 大数据交流群： 376932160 * * * * * ***/ public class HaoopChain { /** * 需求： * 在第一个Mapper里面过滤大于10000万的数据 * 第二个Mapper里面过滤掉大于100-10000的数据 * Reduce里面进行分类汇总并输出 * Reduce后的Mapper里过滤掉商品名长度大于3的数据 */ /** * * 过滤掉大于10000万的数据 * * */ private static class AMapper01 extends MapReduceBase implements Mapper<LongWritable, Text, Text, Text>{ @Override public void map(LongWritable key, Text value, OutputCollector<Text, Text> output, Reporter reporter) throws IOException { String text=value.toString(); String texts[]=text.split(" "); System.out.println("AMapper01里面的数据: "+text); if(texts[1]!=null&&texts[1].length()>0){ int count=Integer.parseInt(texts[1]); if(count>10000){ System.out.println("AMapper01过滤掉大于10000数据: "+value.toString()); return; }else{ output.collect(new Text(texts[0]), new Text(texts[1])); } } } } /** * * 过滤掉大于100-10000的数据 * * */ private static class AMapper02 extends MapReduceBase implements Mapper<Text, Text, Text, Text>{ @Override public void map(Text key, Text value, OutputCollector<Text, Text> output, Reporter reporter) throws IOException { int count=Integer.parseInt(value.toString()); if(count>=100&&count<=10000){ System.out.println("AMapper02过滤掉的小于10000大于100的数据: "+key+" "+value); return; } else{ output.collect(key, value); } } } /** * Reuduce里面对同种商品的 * 数量相加数据即可 * * **/ private static class AReducer03 extends MapReduceBase implements Reducer<Text, Text, Text, Text>{ @Override public void reduce(Text key, Iterator<Text> values, OutputCollector<Text, Text> output, Reporter reporter) throws IOException { int sum=0; System.out.println("进到Reduce里了"); while(values.hasNext()){ Text t=values.next(); sum+=Integer.parseInt(t.toString()); } //旧API的集合，不支持foreach迭代 // for(Text t:values){ // sum+=Integer.parseInt(t.toString()); // } output.collect(key, new Text(sum+"")); } } /*** * * Reduce之后的Mapper过滤 * 过滤掉长度大于3的商品名 * * **/ private static class AMapper04 extends MapReduceBase implements Mapper<Text, Text, Text, Text>{ @Override public void map(Text key, Text value, OutputCollector<Text, Text> output, Reporter reporter) throws IOException { int len=key.toString().trim().length(); if(len>=3){ System.out.println("Reduce后的Mapper过滤掉长度大于3的商品名： "+ key.toString()+" "+value.toString()); return ; }else{ output.collect(key, value); } } } /*** * 驱动主类 * **/ public static void main(String[] args) throws Exception{ //Job job=new Job(conf,"myjoin"); JobConf conf=new JobConf(HaoopChain.class); conf.set("mapred.job.tracker","192.168.75.130:9001"); conf.setJobName("t7"); conf.setJar("tt.jar"); conf.setJarByClass(HaoopChain.class); // Job job=new Job(conf, "2222222"); // job.setJarByClass(HaoopChain.class); System.out.println("模式： "+conf.get("mapred.job.tracker"));; // job.setMapOutputKeyClass(Text.class); // job.setMapOutputValueClass(Text.class); //Map1的过滤 JobConf mapA01=new JobConf(false); ChainMapper.addMapper(conf, AMapper01.class, LongWritable.class, Text.class, Text.class, Text.class, false, mapA01); //Map2的过滤 JobConf mapA02=new JobConf(false); ChainMapper.addMapper(conf, AMapper02.class, Text.class, Text.class, Text.class, Text.class, false, mapA02); //设置Reduce JobConf recduceFinallyConf=new JobConf(false); ChainReducer.setReducer(conf, AReducer03.class, Text.class, Text.class, Text.class, Text.class, false, recduceFinallyConf); //Reduce过后的Mapper过滤 JobConf reduceA01=new JobConf(false); ChainReducer.addMapper(conf, AMapper04.class, Text.class, Text.class, Text.class, Text.class, true, reduceA01); conf.setOutputKeyClass(Text.class); conf.setOutputValueClass(Text.class); conf.setInputFormat(org.apache.hadoop.mapred.TextInputFormat.class); conf.setOutputFormat(org.apache.hadoop.mapred.TextOutputFormat.class); FileSystem fs=FileSystem.get(conf); // Path op=new Path("hdfs://192.168.75.130:9000/root/outputchain"); if(fs.exists(op)){ fs.delete(op, true); System.out.println("存在此输出路径，已删除！！！"); } // // org.apache.hadoop.mapred.FileInputFormat.setInputPaths(conf, new Path("hdfs://192.168.75.130:9000/root/inputchain")); org.apache.hadoop.mapred.FileOutputFormat.setOutputPath(conf, op); // //System.exit(conf.waitForCompletion(true)?0:1); JobClient.runJob(conf); } }
运行日志如下：
模式： 192.168.75.130:9001 存在此输出路径，已删除！！！ WARN - JobClient.copyAndConfigureFiles(746) | Use GenericOptionsParser for parsing the arguments. Applications should implement Tool for the same. WARN - NativeCodeLoader.<clinit>(52) | Unable to load native-hadoop library for your platform... using builtin-java classes where applicable WARN - LoadSnappy.<clinit>(46) | Snappy native library not loaded INFO - FileInputFormat.listStatus(199) | Total input paths to process : 1 INFO - JobClient.monitorAndPrintJob(1380) | Running job: job_201405072054_0009 INFO - JobClient.monitorAndPrintJob(1393) | map 0% reduce 0% INFO - JobClient.monitorAndPrintJob(1393) | map 50% reduce 0% INFO - JobClient.monitorAndPrintJob(1393) | map 100% reduce 0% INFO - JobClient.monitorAndPrintJob(1393) | map 100% reduce 33% INFO - JobClient.monitorAndPrintJob(1393) | map 100% reduce 100% INFO - JobClient.monitorAndPrintJob(1448) | Job complete: job_201405072054_0009 INFO - Counters.log(585) | Counters: 30 INFO - Counters.log(587) | Job Counters INFO - Counters.log(589) | Launched reduce tasks=1 INFO - Counters.log(589) | SLOTS_MILLIS_MAPS=11357 INFO - Counters.log(589) | Total time spent by all reduces waiting after reserving slots (ms)=0 INFO - Counters.log(589) | Total time spent by all maps waiting after reserving slots (ms)=0 INFO - Counters.log(589) | Launched map tasks=2 INFO - Counters.log(589) | Data-local map tasks=2 INFO - Counters.log(589) | SLOTS_MILLIS_REDUCES=9972 INFO - Counters.log(587) | File Input Format Counters INFO - Counters.log(589) | Bytes Read=183 INFO - Counters.log(587) | File Output Format Counters INFO - Counters.log(589) | Bytes Written=19 INFO - Counters.log(587) | FileSystemCounters INFO - Counters.log(589) | FILE_BYTES_READ=57 INFO - Counters.log(589) | HDFS_BYTES_READ=391 INFO - Counters.log(589) | FILE_BYTES_WRITTEN=174859 INFO - Counters.log(589) | HDFS_BYTES_WRITTEN=19 INFO - Counters.log(587) | Map-Reduce Framework INFO - Counters.log(589) | Map output materialized bytes=63 INFO - Counters.log(589) | Map input records=10 INFO - Counters.log(589) | Reduce shuffle bytes=63 INFO - Counters.log(589) | Spilled Records=8 INFO - Counters.log(589) | Map output bytes=43 INFO - Counters.log(589) | Total committed heap usage (bytes)=336338944 INFO - Counters.log(589) | CPU time spent (ms)=1940 INFO - Counters.log(589) | Map input bytes=122 INFO - Counters.log(589) | SPLIT_RAW_BYTES=208 INFO - Counters.log(589) | Combine input records=0 INFO - Counters.log(589) | Reduce input records=4 INFO - Counters.log(589) | Reduce input groups=3 INFO - Counters.log(589) | Combine output records=0 INFO - Counters.log(589) | Physical memory (bytes) snapshot=460980224 INFO - Counters.log(589) | Reduce output records=2 INFO - Counters.log(589) | Virtual memory (bytes) snapshot=2184105984 INFO - Counters.log(589) | Map output records=4
总结，测试过程中，发现如果Reduce后面，还有Mapper执行，那么注意一定要，在ChainReducer里面先set一个全局唯一的Reducer，然后再add一个Mapper，否则，在运行的时候，会报空指针异常，这一点需要特别注意！
Hadoop的ChainMapper和ChainReducer实战,布布扣,bubuko.com
Hadoop的ChainMapper和ChainReducer实战
原文：http://www.cnblogs.com/netskill/p/3914394.html