python hadoop stream 传参

用python写mapreduce，会遇到通过在shell脚本中传参数，通常有两种方式

1.第一种方式通过sys.argv 获取参数值

mapper代码: countmapper.py

import sys

arg1 = sys.argv[1]
arg2 = sys.argv[2]

for line in sys.stdin:
  line = line.strip()
  item, count = line.split(',')
  print '%s\t%s' % (item, count)

说明sys.argv获取参数下标是从1开始

命令配置方式

> hadoop jar hadoop-streaming.jar \
> -mapper 'count_mapper.py arg1 arg2' -file count_mapper.py \
> -reducer 'count_reducer.py arg3' -file count_reducer.py \
> ...

2.第二种方式通过os.environ.get获取参数值

mapper代码:

#!/usr/bin/env python  
# vim: set fileencoding=utf-8  
import sys  
import os  
  
  
def main():  
    card_start = os.environ.get('card_start')  
    card_last = os.environ.get('card_last')  
    trans_at = float(os.environ.get('trans_at'))  
  
    for line in sys.stdin:  
        detail = line.strip().split(',')  
        card = detail[0]  
        money = float(detail[17])  
        if trans_at == money and card_start == card[1 : 7] and card_last == card[-4 : ]:  
            print '%s\t%s' % (line.strip(), detail[1])  

通过参数名称获取值

命令配置方式

hadoop jar ./hadoop-streaming-2.0.0-mr1-cdh4.7.0.jar \  
        -input $1 \  
        -output trans_record/result \  
        -file map.py \  
        -file reduce.py \  
        -mapper "python map.py" \  
        -reducer "python reduce.py" \  
        -jobconf mapred.reduce.tasks=1 \  
        -jobconf mapred.job.name="qianjc_trans_record" \  
        -cmdenv "card_start=$2" \  
        -cmdenv "card_last=$3" \  
        -cmdenv "trans_at=$4" 

个人微信公众号