1.1. 前言
我们计算URL请求数经过这四个步骤:
Mapper: 将以行数据解析成 key=请求的URL数 value=1的形式文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
Shuffle: 通过Shuffle后的结果会生成以 key 的值排序的 value迭代器文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
结果如: 请求的URL数 [1, 1, 1 ... 1, 1]文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
Reduce 1: 在这边我们计算出 每个请求的URL 的访问量文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
输出如: None [sum([1, 1, 1 ... 1, 1]), key]文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
Reduce 2: 对sum([1, 1, 1 ... 1, 1]) 进行排序并输出 TOP 100文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
输入如: 246361 "/wp-admin/admin-ajax.php"文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
1.2. 代码
cat mr_url_req.py
# -*- coding: utf-8 -*-
from mrjob.job import MRJob
from mrjob.step import MRStep
from ng_line_parser import NgLineParser
import heapq
class MRUrlRef(MRJob):
ng_line_parser = NgLineParser()
def mapper(self, _, line):
self.ng_line_parser.parse(line)
yield self.ng_line_parser.reference_url, 1 # 外链域名
def reducer_sum(self, key, values):
"""统计 VU"""
yield None, [sum(values), key]
def reducer_top100(self, _, values):
"""访问数降序"""
for cnt, path in heapq.nlargest(100, values):
yield cnt, path
def steps(self):
return [
MRStep(mapper=self.mapper,
reducer=self.reducer_sum),
MRStep(reducer=self.reducer_top100)
]
def main():
MRUrlRef.run()
if __name__ == '__main__':
main()
运行统计和输出结果文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
python mr_url_req.py < www.ttmark.com.access.log No configs found; falling back on auto-configuration Creating temp directory /tmp/mr_url_req.root.20160924.133027.483660 Running step 1 of 2... reading from STDIN Streaming final output from /tmp/mr_url_req.root.20160924.133027.483660/output... 246361 "/wp-admin/admin-ajax.php" 126012 "/tag/" 57325 "/" ...... 778 "/meirong/2016/03/15/8442.html" 776 "/jiaju/2015/05/30/6058.html" 773 "/jiaju/2015/05/15/5747.html" Removing temp directory /tmp/mr_url_req.root.20160924.133027.483660...
值得一提的是我们在这边使用了 heapq.nlargest 函数来计算 TOP 100 的URL请求数，或许多的程序员估计会使用 sorted([1, 1, 1 ... 1, 1])[:100] 这中形式来做。个人建议在数据量大的时候千万别这么高这样内存会被吃尽并报OOM错误。主要是因为 sorted 先生成了一个list 再对list进行切片。如果list数据量大就有问题。文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
昵称: HH文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
QQ: 275258836文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
ttlsa群交流沟通(QQ群②: 6690706 QQ群③: 168085569 QQ群④: 415230207（新） 微信公众号: ttlsacom)文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
感觉本文内容不错，读后有收获？文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
逛逛衣服店，鼓励作者写出更好文章。文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/ 文章源自运维生存时间-https://www.ttlsa.com/python/python-big-data-analysis-url-req-count-mrjob/
评论