ODPS MR开发 WordCount

本文主要是介绍ODPS MR开发 WordCount，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

参考：
ODPS初始篇--客户端配置和使用：http://blog.itpub.net/26613085/viewspace-1327313/
odps dship客户端使用：http://blog.itpub.net/26613085/viewspace-1328434/
有了上面两篇文章，就可以使用ODPS的客户端；使用ODPS DSHIP往ODPS上上传数据。

1、在Eclipse中创建一个JAVA工程ODPS_WORD_COUNT
2、下载ODPS JAVA SDK包，http://www.aliyun.com/product/odps/，“开发者资源”，“JAVA SDK包”，解压缩。把这些jar包搜罗到一个文件夹下吧，等下添加的时候比较省事。

3、在工程中添加ODPS SDK的JAR包，需要将这些JAR文件添加到Eclipse Project的Build Path中：选中ODPS_WORD_COUNT，右键Bulid Path,Configure Build Path,Java Build Path,Libraries,Add External JARs,选中JAR，然后OK就可以。把上面JAVA SDK包中的Jar文件都添加进来。（如果想要在本地测试MR程序，需要把ODPS CLT里面的jar包都添加到项目中）
4、在ODPS中准备一张表，yangsw_test.word_count.
odps:sql:yangsw_test> create table word_count(content string);
InstanceId: 2014111201412840gqtigdx5
OK
odps:sql:yangsw_test> desc word_count;
+------------------------------------------------------------------------------------+
| Table: word_count                                                                  |
| Owner: ALIYUN$******@***.com     | Project: yangsw_test                       |
| TableComment:                                                                      |
+------------------------------------------------------------------------------------+
| CreatedTime:              2014-11-12 09:41:28                                      |
| LastMetaModifiedTime:     2014-11-12 09:41:28                                      |
| LastDataModifiedTime:     1970-01-01 08:00:00                                      |
+------------------------------------------------------------------------------------+
| Type : Table                 | Size: 0 Bytes                                       |
+------------------------------------------------------------------------------------+
| Native Columns:                                                                    |
+------------------------------------------------------------------------------------+
| Field           | Type       | Comment                                             |
+------------------------------------------------------------------------------------+
| content         | STRING     |                                                     |
+------------------------------------------------------------------------------------+

5、准备数据文件word_count.txt
hello word
hello odps
let us begin cloud compute now
6、 dship导入到word_count表
dship upload -fd "~" -rd "\r\n" C:\Users\yangswa\Desktop\word_count.txt word_count
Upload session: 201411120951352581870a0025cfcc
2014-11-12 09:51:06     scanning file: 'word_count.txt'
2014-11-12 09:51:06     uploading file: 'word_count.txt'
2014-11-12 09:51:07     'word_count.txt' uploaded
OK

7、查看下这张表的数据
odps:sql:yangsw_test> select * from word_count;
InstanceId: 2014111214471814gabmgdx5
SQL: .
+---------+
| content |
+---------+
| hello word |
| hello odps |
| let us begin cloud compute now |
+---------+
8、再创建一个结果表
odps:sql:yangsw_test> create table word_count_result(word string,count bigint);
InstanceId: 20141112074523943g9l5pdx5
OK
odps:sql:yangsw_test> desc word_count_result;
+------------------------------------------------------------------------------------+
| Table: word_count_result                                                           |
| Owner: ALIYUN$**********@126.com     | Project: yangsw_test                       |
| TableComment:                                                                      |
+------------------------------------------------------------------------------------+
| CreatedTime:              2014-11-12 15:45:24                                      |
| LastMetaModifiedTime:     2014-11-12 15:45:24                                      |
| LastDataModifiedTime:     1970-01-01 08:00:00                                      |
+------------------------------------------------------------------------------------+
| Type : Table                 | Size: 0 Bytes                                       |
+------------------------------------------------------------------------------------+
| Native Columns:                                                                    |
+------------------------------------------------------------------------------------+
| Field           | Type       | Comment                                             |
+------------------------------------------------------------------------------------+
| word            | STRING     |                                                     |
| count           | BIGINT     |                                                     |
+------------------------------------------------------------------------------------+
9、实现程序
先建立一个包吧，com.thomas.odps
先借用官方的一个WordCount.java代码，下一篇文章将这个类改写成mapper、reducer和main类分开的写法。这次先把流程都走通。
package com.thomas.odps;

import java.io.IOException;
import java.util.Iterator;
import com.aliyun.odps.Column;
import com.aliyun.odps.OdpsException;
import com.aliyun.odps.OdpsType;
import com.aliyun.odps.data.Record;
import com.aliyun.odps.data.TableInfo;
import com.aliyun.odps.mapred.JobClient;
import com.aliyun.odps.mapred.MapperBase;
import com.aliyun.odps.mapred.ReducerBase;
import com.aliyun.odps.mapred.RunningJob;
import com.aliyun.odps.mapred.conf.JobConf;
import com.aliyun.odps.mapred.utils.InputUtils;
import com.aliyun.odps.mapred.utils.OutputUtils;

public class WordCount {

public static class TokenizerMapper extends MapperBase {
Record word;
Record one;

  @Override
  public void setup(TaskContext context) throws IOException {
   word = context.createMapOutputKeyRecord();
   one = context.createMapOutputValueRecord();
   one.setBigint(0, 1L);
  }

  @Override
  public void map(long recordNum, Record record, TaskContext context)
    throws IOException {
   for (int i = 0; i < record.getColumnCount(); i++) {
    String[] words = record.get(i).toString().split("\\s+");
    for (String w : words) {
     word.setString(0, w);
     context.write(word, one);
    }
   }
  }

public static class SumCombiner extends ReducerBase {
private Record count;

   @Override
   public void setup(TaskContext context) throws IOException {
    count = context.createMapOutputValueRecord();
   }

   @Override
   public void reduce(Record key, Iterator<Record> values,
     TaskContext context) throws IOException {
    long c = 0;
    while (values.hasNext()) {
     Record val = values.next();
     c += (Long) val.get(0);
    }
    count.set(0, c);
    context.write(key, count);
   }
  }

public static class SumReducer extends ReducerBase {
private Record result;

   @Override
   public void setup(TaskContext context) throws IOException {
    result = context.createOutputRecord();
   }

   @Override
   public void reduce(Record key, Iterator<Record> values,
     TaskContext context) throws IOException {
    long count = 0;
    while (values.hasNext()) {
     Record val = values.next();
     count += (Long) val.get(0);
    }
    result.set(0, key.get(0));
    result.set(1, count);
    context.write(result);
   }
  }

  public static void main(String[] args) throws OdpsException {
   if (args.length != 2) {
    System.err.println("Usage: wordcount <in_table> <out_table>");
    System.exit(2);
   }
   JobConf job = new JobConf();
   job.setMapperClass(TokenizerMapper.class);
   job.setCombinerClass(SumCombiner.class);
   job.setReducerClass(SumReducer.class);
   job.setMapOutputKeySchema(new Column[] { new Column("word",
     OdpsType.STRING) });
   job.setMapOutputValueSchema(new Column[] { new Column("count",
     OdpsType.BIGINT) });
   InputUtils.addTable(TableInfo.builder().tableName(args[0]).build(), job);
   OutputUtils.addTable(TableInfo.builder().tableName(args[1]).build(), job);
   RunningJob rj = JobClient.runJob(job);
   rj.waitForCompletion();
  }
}
}
10、本地测试
WordCount.java右键，“Run As”，“Run Configurations …”：

“Main”,”Project”:
ODPS_WORD_COUNT
“Main”,”Main class”:
com.thomas.odps.WordCount$TokenizerMapper
(这个本来应该是com.thomas.odps.WordCount但是系统自动带出来上面的，改掉在本地执行时还报错因此就设置成上面的吧)
“Arguments”，”Program arguments”:
word_count word_count_result
“Arguments”，”VM arguments”:
-Dodps.runner.mode=local
-Dodps.project.name=yangsw_test
-Dodps.end.point=http://service.odps.aliyun.com/api
-Dodps.access.id=**********
-Dodps.access.key=**********
配置好后点击“Run”，启动本地测试

控制台输出：
Running open mr on old console.
2014-11-13 16:15:35 com.aliyun.odps.mapred.LocalJobRunner submit
信息: run mapreduce job in local mode
2014-11-13 16:15:35 com.aliyun.odps.mapred.LocalJobRunner submit
信息: job id: mr_20141113161535_909_4984
2014-11-13 16:15:35 com.aliyun.odps.mapred.LocalJobRunner processInput
信息: Processing input: word_count
2014-11-13 16:15:36 com.aliyun.odps.mapred.local.utils.SchemaUtils generateSchemaFile
信息: generate schema file: D:\MyEclipseWorkStation\ODPS_WORD_COUNT\warehouse\yangsw_test\word_count\__schema__
2014-11-13 16:15:36 com.aliyun.odps.mapred.local.utils.LocalRunUtils downloadTableSchemeAndData
信息: Create table scheme of word_count, Path: D:\MyEclipseWorkStation\ODPS_WORD_COUNT\warehouse\yangsw_test\word_count
2014-11-13 16:15:36 com.aliyun.odps.mapred.local.utils.LocalRunUtils downloadTable
信息: Start to download table: yangsw_test.word_count partSpec: null to D:\MyEclipseWorkStation\ODPS_WORD_COUNT\warehouse\yangsw_test\word_count
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.utils.SchemaUtils generateSchemaFile
信息: generate schema file: D:\MyEclipseWorkStation\ODPS_WORD_COUNT\temp\mr_20141113161535_909_4984\input\yangsw_test\word_count\__schema__
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.utils.SchemaUtils generateSchemaFile
信息: generate schema file: D:\MyEclipseWorkStation\ODPS_WORD_COUNT\temp\mr_20141113161535_909_4984\output\__default__\__schema__
2014-11-13 16:15:37 com.aliyun.odps.mapred.LocalJobRunner runJob
信息: Start to run mappers, num: 1
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver <init>
信息: Map M_000000: input: temp\mr_20141113161535_909_4984\input\yangsw_test\word_count\data:0+56
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver <init>
信息: create record reader: finished
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver run
信息: Start to run Mapper: M_000000
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver$ProxiedMapContextImpl close
信息: Start to run Combiner
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver$ProxiedMapContextImpl close
信息: Fininshed run Combiner
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.MapDriver run
信息: Fininshed run Reducer: M_000000
2014-11-13 16:15:37 com.aliyun.odps.mapred.LocalJobRunner runJob
信息: Start to run reduces, num: 1
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.ReduceDriver run
信息: Start to run Reducer: R_000000
2014-11-13 16:15:37 com.aliyun.odps.mapred.local.ReduceDriver run
信息: Fininshed run Reducer: R_000000
2014-11-13 16:15:37 com.aliyun.odps.mapred.LocalJobRunner moveOutputs
信息: Copy output to warehouse: label=__default__ -> D:\MyEclipseWorkStation\ODPS_WORD_COUNT\warehouse\yangsw_test\word_count_result
Summary:
counters: 8
map-reduce framework
  combine_input_groups=9
  combine_output_records=9
  map_input_bytes=56
  map_input_records=3
  map_output_records=10
  reduce_output_[word_count_result]_bytes=74
  reduce_output_[word_count_result]_records=9

OK
InstanceId: mr_20141113161535_909_4984

执行完成后，会在该工程的目录下面多了如下目录结构：
./temp
./ warehouse
/ yangsw_test
    /word_count
        / __schema__
        / data
    /word_count_result
        / __schema__
        / R_000000
其中word_count下面的__schema__内容包含word_count源表的项目和表结构：
project=yangsw_test
table=word_count
columns=content:STRING
word_count下面的data是word_count表的数据：
hello word
hello odps
let us begin cloud compute now
word_count_result下面的__schema__是目标表的项目和表结构：
project=yangsw_test
table=word_count_result
columns=word:STRING,count:BIGINT
word_count_result下面的R_000000是输出结果，本次只有一个reduce所以只有一个文件：
begin,1
cloud,1
compute,1
hello,2
let,1
now,1
odps,1
us,1
word,1
上面就是wordcount本地运行的结果了
11、发布jar包
create resource jar C:\Users\yangswa\Desktop\word_count.jar word_count.jar
12、在服务端运行
odps:yangsw_test> jar -resources word_count.jar -classpath C:\Users\yangswa\Desktop\word_count.jar com.thomas.odps.WordCount
word_count word_count_result
Exception in thread "main" java.lang.RuntimeException: java.lang.NoSuchMethodException: com.thomas.odps.WordCount.main([Ljava
.lang.String;)
        at com.aliyun.odps.mapred.cli.Cli.main(Cli.java:47)
Caused by: java.lang.NoSuchMethodException: com.thomas.odps.WordCount.main([Ljava.lang.String;)
        at java.lang.Class.getMethod(Unknown Source)
        at com.aliyun.odps.mapred.cli.Cli.main(Cli.java:35)
这个居然报错：main方法不存在，在odps用户中心的售后服务中，提交了个工单，得到的反馈是ODPS MR功能还没有正式开放，要等到12月底才能正式使用。好吧，就先等等吧.