hadoop程序开发 --- python

这里以统计单词为例

1 首先建立mapper.py

mkdir /usr/local/hadoop-python
cd /usr/local/hadoop-python
vim mapper.py

mapper.py

#!/usr/bin/env python

import sys

# input comes from STDIN (standard input) 输入来自STDIN（标准输入）
for line in sys.stdin:
    # remove leading and trailing whitespace 删除前导和尾随空格
    line = line.strip()
    # split the line into words 把线分成单词
    words = line.split()
    # increase counters 增加柜台
    for word in words:
        # write the results to STDOUT (standard output); 
        # 将结果写入STDOUT（标准输出）；
        # what we output here will be the input for the
        # Reduce step, i.e. the input for reducer.py
        # tab-delimited; the trivial word count is 1
        # 我们在此处输出的内容将是Reduce步骤的输入，即reducer.py制表符分隔的输入; 		   # 平凡的字数是1
        print '%s\t%s' % (word, 1)

文件保存后，请注意将其权限作出相应修改：

chmod a+x /usr/local/hadoop-python/mapper.py

2 建立reducer.py

vim reducer.py

#!/usr/bin/env python

from operator import itemgetter
import sys

current_word = None
current_count = 0
word = None

# input comes from STDIN 输入来自STDIN
for line in sys.stdin:
    # remove leading and trailing whitespace 
    # 删除前导和尾随空格
    line = line.strip()

    # parse the input we got from mapper.py
    # 解析我们从mapper.py获得的输入
    word, count = line.split('\t', 1)

    # convert count (currently a string) to int
    # 将count（当前为字符串）转换为int
    try:
        count = int(count)
    except ValueError:
        # count was not a number, so silently
        # ignore/discard this line
        # count不是数字，因此请忽略/丢弃此行
        continue

    # this IF-switch only works because Hadoop sorts map output
    # by key (here: word) before it is passed to the reducer
    # 该IF开关仅起作用是因为Hadoop在将映射输出传递给reducer之前按键（此处为word）对	  # 映射输出进行排序
    if current_word == word:
        current_count += count
    else:
        if current_word:
            # write result to STDOUT
            # 将结果写入STDOUT
            print '%s\t%s' % (current_word, current_count)
        current_count = count
        current_word = word

# do not forget to output the last word if needed!
# 如果需要，不要忘记输出最后一个单词！
if current_word == word:
    print '%s\t%s' % (current_word, current_count)

文件保存后，请注意将其权限作出相应修改：

chmod a+x /usr/local/hadoop-python/reducer.py

首先可以在本机上测试以上代码，这样如果有问题可以及时发现：

# echo "foo foo quux labs foo bar quux" | /usr/local/hadoop-python/mapper.py

输出：
foo	1
foo	1
quux	1
labs	1
foo	1
bar	1
quux	1

再运行以下包含reduce.py的代码：

echo "foo foo quux labs foo bar quux" | /usr/local/hadoop-python/mapper.py | sort -k1,1 | /usr/local/hadoop-python/reducer.py

输出：
bar		1
foo		3
labs	1
quux	2

3 在Hadoop上运行Python代码

准备工作：
下载文本文件：

yum install wget -y

mkdir input
cd /usr/local/hadoop-python/input

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

然后把这二本书上传到hdfs文件系统上：

# 在hdfs上的该用户目录下创建一个输入文件的文件夹
hdfs dfs -mkdir /input 

# 上传文档到hdfs上的输入文件夹中
hdfs dfs -put /usr/local/hadoop-python/input/pg20417.txt /input 

寻找你的streaming的jar文件存放地址，注意2.6的版本放到share目录下了，可以进入hadoop安装目录寻找该文件：

cd $HADOOP_HOME
find ./ -name "*streaming*.jar"

然后就会找到我们的share文件夹中的hadoop-straming*.jar文件:

./share/hadoop/tools/lib/hadoop-streaming-2.8.4.jar
./share/hadoop/tools/sources/hadoop-streaming-2.8.4-test-sources.jar
./share/hadoop/tools/sources/hadoop-streaming-2.8.4-sources.jar

/usr/local/hadoop-2.8.4/share/hadoop/tools/lib

由于这个文件的路径比较长，因此我们可以将它写入到环境变量：

vim /etc/profile

export STREAM=/usr/local/hadoop-2.8.4/share/hadoop/tools/lib/hadoop-streaming-2.8.4.jar

由于通过streaming接口运行的脚本太长了，因此直接建立一个shell名称为run.sh来运行：

vim run.sh

hadoop jar /usr/local/hadoop-2.8.4/share/hadoop/tools/lib/hadoop-streaming-2.8.4.jar \
-files /usr/local/hadoop-python/mapper.py,/usr/local/hadoop-python/reducer.py \
-mapper /usr/local/hadoop-python/mapper.py \
-reducer /usr/local/hadoop-python/reducer.py \
-input /input/pg20417.txt \
-output /output1

hadoop jar $STREAM \-files /usr/local/hadoop-python/mapper.py,/usr/local/hadoop-python/reducer.py \-mapper /usr/local/hadoop-python/mapper.py \-reducer /usr/local/hadoop-python/reducer.py \-input /input/pg20417.txt \-output /output1