python脚本实现统计日志文件中的ip访问次数代码分享

适用的日志格式：

106.45.185.214 - - [06/Aug/2014:07:38:59 +0800] "GET / HTTP/1.0" 200 10 "-" "-"
171.104.119.22 - - [06/Aug/2014:08:55:01 +0800] "GET / HTTP/1.0" 200 10 "-" "-"
27.31.238.242 - - [06/Aug/2014:09:43:19 +0800] "GET / HTTP/1.0" 200 10 "-" "-"
218.65.202.131 - - [06/Aug/2014:10:33:59 +0800] "GET / HTTP/1.0" 200 10 "-" "-"

以上为nginx的日志,本程序只适合用于IP在开头并用空格分开的这种格式。

实例代码：

import sys

class Log:

def __init__(self, filename,dic,count):
   self.filename = filename
   self.dic=dic
   self.count=count

def parse(self):
   i=1
   f=file(self.filename)
   while True:
     line=f.readline()
     if len(line)==0:
       break
     ip=line.split(' ')
     if ip[0] in dic:
       self.dic[ip[0]]=self.dic[ip[0]]+1
     else:
       self.dic[ip[0]]=i
   soredic=sorted(self.dic.items(), key=lambda d:d[1],reverse=True)
   counts=0;
   for item in soredic:
     if counts==int(self.count):
       break
     print("IP:％s  Total Times: ％s"％(item[0],item[1]))
     counts=counts+1
   f.close()

if __name__=="__main__":

if len(sys.argv)<3:
   print('usage:log.py log.log toptimes\nexample log.py log.log 20\ncode by iswin')
   sys.exit()
 dic={}
 log=Log(sys.argv[1],dic,sys.argv[2])
 log.parse()