记录第一次写python(查找日志里重复的条码数量)

这个python的代码，是用来查找日志里，记录的条码的重复数量。

#!/usr/bin/python
# -*- coding: UTF-8 -*-

#目前的字典会自动过滤【时间】完全一致的数据，需要注意

import datetime
import os
import sys
import math

reload(sys)
sys.setdefaultencoding('utf8')

def view_bar(Index):
    Index="%02d" % Index  #数字前面补齐0
    sys.stdout.write(Index+'%->'+"\b\b\b\b\b")
    sys.stdout.flush()
    
def Remove0D0A(tmpstr):
    return tmpstr.strip().replace('\n', '').replace('\r', '')

def GetAllBarcode(filename):
    searchstr='Applicator:'
    l=len(searchstr)
    pos ={}
    with open(filename,'r') as file_to_read:
        while True:
            line = file_to_read.readline()
            if not line :
                break
            pass
            if line.find('\''+searchstr+'\'')>0:
                continue
            tindex=line.find('\t')
            bindex=line.find(searchstr)
            if bindex>0:
                time=''
                tmp=line[bindex+l:]
                if tindex>0:
                    time=line[0:tindex]
                else:
                    time=tmp       #如果时间不存在，则以条码记录作为时间列                    
                    global Count
                    Count=Count+1
                tmp=line[bindex+l:]
                pos[time]=tmp
                #print 'Time:',time,'Barcode:',tmp
    #print pos
    return pos

Count=0
start = datetime.datetime.now()
dir='DataFile'
files=os.listdir(dir)
print 'Analyze File Count:',len(files)
TimeBarcode={}
for a in files:
    path=dir+'\\'+a
    #print 'Path:',path
    tmp=GetAllBarcode(path)
    TimeBarcode.update(tmp)
    #print tmp

Barcode=[]
f = open('record.csv','a+')
for time,value in TimeBarcode.items():
    #print 'Time:',time,'Barcode:',value
    time=Remove0D0A(time)
    barcode=Remove0D0A(value)
    Barcode.append(barcode)
    f.write(time+',\''+barcode+'\r')
print 'Barcode Total Count:',len(Barcode)
read= datetime.datetime.now()
print 'Read use time:',str(read-start)

duplicated=set()
barcode_len=len(Barcode)
LastNum=0
for i in range(0,barcode_len):
    Num=(float(i)/barcode_len)*100
    if math.modf(Num)[1]-LastNum>0:
        LastNum=Num
        view_bar(LastNum)
    if(Barcode[i] in Barcode[i+1:]):
        duplicated.add(Barcode[i])
    elif(i>1 and Barcode[i] in Barcode[0:i-1]):
        duplicated.add(Barcode[i],)

if len(duplicated)>0:
    f = open('duplicate.csv','a+')
    for b in duplicated:
        f.write('\''+b+'\r')
        print b+'\r'
    #print duplicated
else:
 print 'No duplicate'

end= datetime.datetime.now()
print 'Analyze data use time:',str(end-read)
if Count>0:
    print 'Time not exist:',Count