python对多个同格式文件的某列读取到另外一个文件

#!/usr/bin/python
# -*- coding: UTF-8 -*-
import os
import glob
import sys
import re
import datetime
# 设置编码
reload(sys)
sys.setdefaultencoding('utf8')

filePath="/home/rfdev/cbh"
#proj1 = "*appStat.log*"

# 默认是读取昨天的日志
today = (datetime.datetime.now()+datetime.timedelta(days=-1)).strftime("%Y%m%d")

todayDb = (datetime.datetime.now()+datetime.timedelta(days=-1)).strftime("%Y-%m-%d")
tomorrowDb = (datetime.datetime.now()).strftime("%Y-%m-%d")

if len(sys.argv) == 2:
today = re.sub("-","",sys.argv[1])
todayDb = sys.argv[1]
tomorrowDb = (datetime.datetime.strptime(todayDb,'%Y-%m-%d')+datetime.timedelta(days=1)).strftime("%Y-%m-%d")

with open(r"/home/rfdev/cbh/test11.txt.tmp",'w') as file_object:
file_object.truncate()

os.chdir(filePath)
logPattern = "*appStat.log." + todayDb
for logFile in glob.iglob(logPattern):
f = open(logFile,"r")
line=f.readline()
while line:
SplashActivity = line.find("pageName=SplashActivity")
PadFragment = line.find("pageName=PadFragment")
RegisterActivity = line.find("pageName=RegisterActivity")
try:
if SplashActivity>0:
SplashActivity_id=line.split("cuid=")[1].split()[0].replace("\x00","")
with open(r"/home/rfdev/cbh/test11.txt.tmp",'a') as file_object:
file_object.write("1,"+SplashActivity_id+"\n")
elif PadFragment>0:
if line.find("date")>0:
if line.find("userId=0")>0:
PadFragment_id=line.split("cuid=")[1].split()[0].replace("\x00","")
with open(r"/home/rfdev/cbh/test11.txt.tmp",'a') as file_object:
file_object.write("2,"+PadFragment_id+"\n")
elif RegisterActivity>0:
RegisterActivity_id=line.split("cuid=")[1].split()[0].replace("\x00","")
with open(r"/home/rfdev/cbh/test11.txt.tmp",'a') as file_object:
file_object.write("3,"+RegisterActivity_id+"\n")
except IndexError,e:
print e.message
print line
line=f.readline()

with open(r"/home/rfdev/cbh/test11_new.txt",'w') as file_object1:
file_object1.truncate()

f1=open('/home/rfdev/cbh/test11.txt.tmp','r')
lines=f1.readline()
while lines:
if len(lines)<500:
with open(r"/home/rfdev/cbh/test11_new.txt",'a') as file_object1:
file_object1.write(lines)
lines=f1.readline()

 

posted @ 2018-01-05 11:45  我是东风破  阅读(1003)  评论(0编辑  收藏  举报