理解MapReduce计算构架

用Python编写WordCount程序任务

程序	WordCount
输入	一个包含大量单词的文本文件
输出	文件中每个单词及其出现次数（频数），并按照单词字母顺序排序，每个单词和其频数占一行，单词和频数之间有间隔

 

1.编写map函数，reduce函数

　　首先在/home/hadoop路径下建立wc文件夹，在wc文件夹下创建文件mapper.py和reducer.py

1 2 3 4

cd /home/hadoop mkdir wc cd /home/hadoop/wc touch mapper.py

1	touch reducer.py

　　

　　编写两个函数

　　mapper.py:

1 2 3 4 5 6 7

#!/usr/bin/env python import sys for line in sys.stdin:     line = line.strip()     words = line.split()     for word in words:         print '%s\t%s' % (word,1)

　　

　　reducer.py:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

#!/usr/bin/env python from operator import itemgetter import sys   current_word = None current_count = 0 word=None   for line in sys.stdin:     line = line.strip()     word, count = line.split('\t', 1)     try:         count=int(count)     except ValueError:         continue       if current_word == word:         current_count += count     else:         if current_word:             print '%s\t%s' % (current_word,  current_count)         current_count = count         current_word = word if current_word == word:     print '%s\t%s' % (current_word,  current_count)

　　

2.将其权限作出相应修改

1 2	chmod a+x /home/hadoop/wc/mapper.py chmod a+x /home/hadoop/wc/reducer.py

 

3.本机上测试运行代码

1 2 3

echo "foo foo quux labs foo bar quux" | /home/hadoop/wc/mapper.py   echo "foo foo quux labs foo bar quux" | /home/hadoop/wc/mapper.py | sort -k1,1 | /home/hadoop/wc/reducer.py

　　

 

4.放到HDFS上运行

　　下载文本文件或爬取网页内容存成的文本文件：

1 2 3

cd  /home/hadoop/wc wget http://www.gutenberg.org/files/5000/5000-8.txt wget http://www.gutenberg.org/cache/epub/20417/pg20417.txt

　　

5.下载并上传文件到hdfs上

1	hdfs dfs -put /home/hadoop/hadoop/gutenberg/*.txt /user/hadoop/input

 

6.用Hadoop Streaming命令提交任务

 　　寻找你的streaming的jar文件存放地址：

1	cd /usr/local/hadoop/share/hadoop/tools/lib/hadoop-streaming-2.7.1.jar

　　打开环境变量配置文件:

1	gedit ~/.bashrc

　　在里面写入streaming路径:

1	export STREAM=$HADOOP_HOME/share/hadoop/tools/lib/hadoop-streaming-*.jar

　　让环境变量生效:

1 2	source ~/.bashrc echo $STREAM

　　建立一个shell名称为run.sh来运行：

1	gedit run.sh

1 2 3 4 5 6 7

hadoop jar $STREAM -file /home/hadoop/wc/mapper.py \ -mapper /home/hadoop/wc/mapper.py \ -file /home/hadoop/wc/reducer.py \ -reducer /home/hadoop/wc/reducer.py \ -input /user/hadoop/input/*.txt \ -output /user/hadoop/wcoutput

1	source run.sh