python批量转换文件编码格式

自己写的方法,适用于linux,

#!/usr/bin/python
#coding=utf-8
import sys
import os, os.path
import dircache
import commands
def add(x,y):
return x*y
def trans(dirname):
lis = dircache.opendir(dirname)
for a in lis:
af=dirname+os.sep+a
## print af
if os.path.isdir(af):
## print af
trans(af)
else:
## print af+”encoding=”+fi.name
ft = commands.getoutput(‘file -i ‘+af)
## print ft
if a.find(‘.htm’)==-1 and a.find(‘.xml’)==-1 and ft.find(‘text/’)!=-1 and ft.find(‘iso-8859’)!=-1:
print ‘gbk’+ft+”>”+af
commands.getoutput(‘iconv -ficonv -f gbk -t utf-8 -c -o’+””+af+””+af)
trans(os.getcwd())

py2.6以下版本可用代码

import os,sys
def convert( filename, in_enc = “gbk”, out_enc=”utf8″ ):
try:
print “convert ” + filename,
content = open(filename).read()
new_content = content.decode(in_enc).encode(out_enc)
open(filename, ‘w’).write(new_content)
print ” done”
except:
print ” error”
def explore(dir):
for root, dirs, files in os.walk(dir):
for file in files:
path = os.path.join(root, file)
convert(path)
def main():
for path in sys.argv[1:]:
if os.path.isfile(path):
convert(path)
elif os.path.isdir(path):
explore(path)
if __name__ == “__main__”:
main()

支持py3.1的版本

import os
import sys
import codecs
#该程序用于将目录下的文件从指定格式转换到指定格式,默认的是gbk转到utf-8
def convert(file,in_enc=”gbk”,out_enc=”utf-8″):
try:
print (“convert ” +file)
f=codecs.open(file,’r’,in_enc)
new_content=f.read()
codecs.open(file,’w’,out_enc).write(new_content)
#print (f.read())
except ioerror as err:
print (“i/o error: {0}”.format(err))
def explore(dir):
for root,dirs,files in os.walk(dir):
for file in files:
path=os.path.join(root,file)
convert(path)
def main():
for path in sys.argv[1:]:
if(os.path.isfile(path)):
convert(path)
elif os.path.isdir(path):
explore(path)
if __name__==”__main__”:
main()

以上所述就是本文 的全部内容了,希望大家能够喜欢。

Posted in 未分类