MapReduce中的自定义多目录/文件名输出HDFS

转载自 http://my.oschina.net/leejun2005/blog/94706

最近考虑到这样一个需求：

需要把原始的日志文件用hadoop做清洗后，按业务线输出到不同的目录下去，以供不同的部门业务线使用。

这个需求需要用到MultipleOutputFormat和MultipleOutputs来实现自定义多目录、文件的输出。

需要注意的是，在hadoop 0.21.x之前和之后的使用方式是不一样的：

hadoop 0.21 之前的API 中有 org.apache.hadoop.mapred.lib.MultipleOutputFormat 和 org.apache.hadoop.mapred.lib.MultipleOutputs，而到了 0.21 之后 的API为 org.apache.hadoop.mapreduce.lib.output.MultipleOutputs ，

新版的API 整合了上面旧API两个的功能，没有了MultipleOutputFormat。

本文将给出新旧两个版本的API code。

1、旧版0.21.x之前的版本：

01 import java.io.IOException;
02
03 import org.apache.hadoop.conf.Configuration;
04 import org.apache.hadoop.conf.Configured;
05 import org.apache.hadoop.fs.Path;
06 import org.apache.hadoop.io.LongWritable;
07 import org.apache.hadoop.io.NullWritable;
08 import org.apache.hadoop.io.Text;
09 import org.apache.hadoop.mapred.FileInputFormat;
10 import org.apache.hadoop.mapred.FileOutputFormat;
11 import org.apache.hadoop.mapred.JobClient;
12 import org.apache.hadoop.mapred.JobConf;
13 import org.apache.hadoop.mapred.MapReduceBase;
14 import org.apache.hadoop.mapred.Mapper;
15 import org.apache.hadoop.mapred.OutputCollector;
16 import org.apache.hadoop.mapred.Reporter;
17 import org.apache.hadoop.mapred.TextInputFormat;
18 import org.apache.hadoop.mapred.lib.MultipleTextOutputFormat;
19 import org.apache.hadoop.util.Tool;
20 import org.apache.hadoop.util.ToolRunner;
21
22 public class MultiFile extendsConfigured implementsTool {
23
24     publicstatic class MapClass extends MapReduceBase implements
25             Mapper<LongWritable, Text, NullWritable, Text> {
26
27         @Override
28         publicvoid map(LongWritable key, Text value,
29                 OutputCollector<NullWritable, Text> output, Reporter reporter)
30                 throwsIOException {
31             output.collect(NullWritable.get(), value);
32         }
33
34     }
35
36     // MultipleTextOutputFormat 继承自MultipleOutputFormat，实现输出文件的分类
37
38     publicstatic class PartitionByCountryMTOF extends
39             MultipleTextOutputFormat<NullWritable, Text> {// key is
40                                                             // NullWritable,
41                                                             // value is Text
42         protectedString generateFileNameForKeyValue(NullWritable key,
43                 Text value, String filename) {
44             String[] arr = value.toString().split(",", -1);
45             String country = arr[4].substring(1,3); // 获取country的名称
46             returncountry + "/" + filename;
47         }
48     }
49
50     // 此处不使用reducer
51     /*
52      * public static class Reducer extends MapReduceBase implements
53      * org.apache.hadoop.mapred.Reducer<LongWritable, Text, NullWritable, Text>
54      * {
55      *
56      * @Override public void reduce(LongWritable key, Iterator<Text> values,
57      * OutputCollector<NullWritable, Text> output, Reporter reporter) throws
58      * IOException { // TODO Auto-generated method stub
59      *
60      * }
61      *
62      * }
63      */
64     @Override
65     publicint run(String[] args)throws Exception {
66         Configuration conf = getConf();
67         JobConf job =new JobConf(conf, MultiFile.class);
68
69         Path in =new Path(args[0]);
70         Path out =new Path(args[1]);
71
72         FileInputFormat.setInputPaths(job, in);
73         FileOutputFormat.setOutputPath(job, out);
74
75         job.setJobName("MultiFile");
76         job.setMapperClass(MapClass.class);
77         job.setInputFormat(TextInputFormat.class);
78         job.setOutputFormat(PartitionByCountryMTOF.class);
79         job.setOutputKeyClass(NullWritable.class);
80         job.setOutputValueClass(Text.class);
81
82         job.setNumReduceTasks(0);
83         JobClient.runJob(job);
84         return0;
85     }
86
87     publicstatic void main(String[] args) throws Exception {
88         intres = ToolRunner.run(newConfiguration(), newMultiFile(), args);
89         System.exit(res);
90     }
91
92 }

测试数据及结果：

1 hadoop fs -cat /tmp/multiTest.txt
2 5765303,1998,14046,1996,"AD","",,1,12,42,5,59,11,1,0.4545,0,0,1,67.3636,,,,
3 5785566,1998,14088,1996,"AD","",,1,9,441,6,69,3,0,1,,0.6667,,4.3333,,,,
4 5894770,1999,14354,1997,"AD","",,1,,82,5,51,4,0,1,,0.625,,7.5,,,,
5 5765303,1998,14046,1996,"CN","",,1,12,42,5,59,11,1,0.4545,0,0,1,67.3636,,,,
6 5785566,1998,14088,1996,"CN","",,1,9,441,6,69,3,0,1,,0.6667,,4.3333,,,,
7 5894770,1999,14354,1997,"CN","",,1,,82,5,51,4,0,1,,0.625,,7.5,,,,

from：

MultipleOutputFormat Example

http://mazd1002.blog.163.com/blog/static/665749652011102553947492/

2、新版0.21.x及之后的版本：

01 public class TestwithMultipleOutputs extendsConfigured implementsTool {
02
03 　　publicstatic class MapClass extends Mapper<LongWritable,Text,Text,IntWritable> {
04
05 　　　　privateMultipleOutputs<Text,IntWritable> mos;
06
07 　　　　protectedvoid setup(Context context)throws IOException,InterruptedException {
08 　　　　　　mos = newMultipleOutputs<Text,IntWritable>(context);
09 　　　　}
10
11 　　　　publicvoid map(LongWritable key, Text value, Context context)throws IOException, InterruptedException{
12 　　　　　　String line = value.toString();
13 　　　　　　String[] tokens = line.split("-");
14
15 　　　　　　mos.write("MOSInt",newText(tokens[0]),new IntWritable(Integer.parseInt(tokens[1]))); //（第一处）
16 　　　　　　mos.write("MOSText",new Text(tokens[0]),tokens[2]);　　　　//（第二处）
17 　　　　　　mos.write("MOSText",new Text(tokens[0]),line,tokens[0]+"/");　　//（第三处）同时也可写到指定的文件或文件夹中
18 　　　　}
19
20 　　　　protectedvoid cleanup(Context context)throws IOException,InterruptedException {
21 　　　　　　mos.close();
22 　　　　}
23
24 　　}
25 　　publicint run(String[] args)throws Exception {
26
27 　　　　Configuration conf = getConf();
28
29 　　　　Job job = newJob(conf,"word count with MultipleOutputs");
30
31 　　　　job.setJarByClass(TestwithMultipleOutputs.class);
32
33 　　　　Path in = newPath(args[0]);
34 　　　　Path out = newPath(args[1]);
35
36 　　　　FileInputFormat.setInputPaths(job, in);
37 　　　　FileOutputFormat.setOutputPath(job, out);
38
39 　　　　job.setMapperClass(MapClass.class);
40 　　　　job.setNumReduceTasks(0);
41
42 　　　　MultipleOutputs.addNamedOutput(job,"MOSInt",TextOutputFormat.class,Text.class,IntWritable.class);
43 　　　　MultipleOutputs.addNamedOutput(job,"MOSText",TextOutputFormat.class,Text.class,Text.class);
44
45 　　　　System.exit(job.waitForCompletion(true)?0:1);
46 　　　　return0;
47 　　}
48
49 　　publicstatic void main(String[] args) throws Exception {
50
51 　　　　intres = ToolRunner.run(newConfiguration(), newTestwithMultipleOutputs(), args);
52 　　　　System.exit(res);
53 　　}
54
55 }

测试的数据：

abc-1232-hdf
abc-123-rtd
ioj-234-grjth
ntg-653-sdgfvd
kju-876-btyun
bhm-530-bhyt
hfter-45642-bhgf
bgrfg-8956-fmgh
jnhdf-8734-adfbgf
ntg-68763-nfhsdf
ntg-98634-dehuy
hfter-84567-drhuk

结果截图：（结果输出到/test/testMOSout）

PS：遇到的一个问题：

　　如果没有mos.close(), 程序运行中会出现异常：

　　12/05/21 20:12:47 WARN hdfs.DFSClient: DataStreamer Exception:

　　org.apache.hadoop.ipc.RemoteException:org.apache.hadoop.hdfs.server.namenode.LeaseExpiredException: No lease on
　　/test/mosreduce/_temporary/_attempt_local_0001_r_000000_0/h-r-00000 File does not exist. [Lease. Holder: DFSClient_-352105532, pendingcreates: 5]

from：

MultipleOutputFormat和MultipleOutputs

http://www.cnblogs.com/liangzh/archive/2012/05/22/2512264.html    

Hadoop利用Partitioner对输出文件分类（改写partition，路由到指定的文件中）

http://superlxw1234.iteye.com/blog/1495465

http://ghost-face.iteye.com/blog/1869926

更多参考&推荐阅读：

1、【Hadoop】利用MultipleOutputs,MultiOutputFormat实现以不同格式输出到多个文件

http://www.cnblogs.com/iDonal/archive/2012/08/07/2626588.html

2、cdh3u3 hadoop 0.20.2 MultipleOutputs 多输出文件初探

http://my.oschina.net/wangjiankui/blog/49521

3、使用MultipleOutputs

http://blog.163.com/ecy_fu/blog/static/444512620101274344951/

4、Hadoop reduce多个输出

http://blog.csdn.net/inte_sleeper/article/details/7042020

5、Hadoop 0.20.2中怎么使用MultipleOutputFormat实现多文件输出和完全自定义文件名

http://www.cnblogs.com/flying5/archive/2011/05/04/2078407.html

6、Hadoop OutputFormat浅析

http://zhb-mccoy.iteye.com/blog/1591635

7、others：

https://sites.google.com/site/hadoopandhive/home/how-to-write-output-to-multiple-named-files-in-hadoop-using-multipletextoutputformat
https://issues.apache.org/jira/browse/HADOOP-3149
http://grokbase.com/t/hadoop/common-user/112ewx7s15/could-i-write-outputs-in-multiple-directories

8、MultipleOutputs 官方范例

http://hadoop.apache.org/docs/stable/api/org/apache/hadoop/mapreduce/lib/output/MultipleOutputs.html

9、多数据源输入：MultipleInputs

http://*.com/questions/17456369/mapreduce-job-with-mixed-data-sources-hbase-table-and-hdfs-files

https://groups.google.com/forum/#!topic/nosql-databases/SH61smOV-mo

http://bigdataprocessing.wordpress.com/2012/07/27/hadoop-hbase-mapreduce-examples/

http://hbase.apache.org/book/mapreduce.example.html