python3.x读写文件及BOM处理详解

>>> s='梯'
# python3.x 字符串 str 对应 unicode
>>> s,type(s)
('梯', <class 'str'>)
>>> s_gbk=s.encode('gbk')
>>> s_utf8=s.encode('utf-8')
# 1个汉字，gbk编码为2个字节， utf-8 编码为3个字节
# 2位16进制 占1个字节
>>> s_gbk,s_utf8
(b'\xcc\xdd', b'\xe6\xa2\xaf')
>>> len(s_gbk),len(s_utf8)
(2, 3)

# 写文件时 encoding 表示文件存储到磁盘的编码格式
>>> open('gbk.txt','w',encoding='gbk').write('梯')
1
>>> open('utf8.txt','w',encoding='utf-8').write('梯')
1
# 'rb' 二进制模式，读取文件内容为bytes类型，为文件实际存储的字节串
# gbk编码，1个中文占2个字节
>>> open('gbk.txt','rb').read()
b'\xcc\xdd'
# utf-8编码，1个中文占3个字节
>>> open('utf8.txt','rb').read()
b'\xe6\xa2\xaf'

# 读文件时 encoding 表示解释器读取文件字节串解码使用的编码格式
# 需与文件存储时编码格式相同
>>> open('gbk.txt','r',encoding='gbk').read()
'梯'
>>> open('utf8.txt','r',encoding='utf-8').read()
'梯'
>>> dgbk=open('gbk.txt','rb').read();dgbk
b'\xcc\xdd'
>>> dutf8=open('utf8.txt','rb').read();dutf8
b'\xe6\xa2\xaf'
# 二进制模式，读取文件内容后进行手动解码
# python3.x 解释器默认编码为utf-8
>>> dgbk.decode('gbk');dutf8.decode()
'梯'
'梯'

import locale
locale.getpreferredencoding()

>>> import locale
>>> locale.getpreferredencoding()
'cp936'
>>> locale.getpreferredencoding(False)
'cp936'

>>> '梯'.encode()
b'\xe6\xa2\xaf'
>>> '梯'.encode('utf-8')
b'\xe6\xa2\xaf'
>>> '梯'.encode('gbk')
b'\xcc\xdd'
>>> '梯'.encode('cp936')
b'\xcc\xdd'
# utf-16 带小端字节序
>>> '梯'.encode('utf-16')
b'\xff\xfe\xafh'
# utf-16 忽略小端字节序(little BOM)
>>> '梯'.encode('utf-16-le')
b'\xafh'

>>> fpath='temp.txt'
# open()函数encoding未传时默认编码取 locale.getpreferredencoding(False)
# 'r+' 文本模式打开文件
>>> f=open(fpath,'r+')
# f.encoding 获取文件对象的编码方式
>>> f.encoding
'cp936'
# 文件指针会停留在文件结尾
>>> f.write('梯')
1
# 文件指针移动到开头
>>> f.seek(0)
0
>>> f.read()
'梯'
>>> f.close()
# 'rb' 二进制模式直接读取文件的字节串
# 与 cp936 编码一致
>>> open(fpath,'rb').read()
b'\xcc\xdd'
# 通过 相同的编码 cp936或gbk 读取文件内容
>>> open(fpath,'r',encoding='cp936').read()
'梯'
>>> open(fpath,'r',encoding='gbk').read()
'梯'
>>> open(fpath,'r',encoding='utf-8').read()
Traceback (most recent call last):
  File "<pyshell#24>", line 1, in <module>
    open(fpath,'r',encoding='utf-8').read()
  File "D:\python3\lib\codecs.py", line 322, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xcc in position 0: invalid continuation byte

# python2.x内置open()函数没有encoding入参，无法指定编码方法
>>> fpath='temp.txt'
>>> f=open(fpath,'w')
>>> f.write('a梯')
>>> f.close()
>>> open(fpath,'r').read()
'a\xcc\xdd'
# 'rb' 二进制模式直接读取文件的字节串
# 与 cp936 编码一致
>>> open(fpath,'rb').read()
'a\xcc\xdd'
>>> len('a梯')
3
# python2.x通过 codecs.open(encoding)指定编码
>>> import codecs
>>> f=codecs.open(fpath,'w','utf-8')
# 写入汉字时需要通过u'汉字'创建Unicode字符串
# 因为python2.x解释器默认编码为ASCII码，
# '梯'字符串str用操作系统cp936自动编码为 '\xcc\xdd'
# 解释器需要对其解码为Unicode，解码时 \xcc 不在ASCII范围内，会报错
>>> f.write(u'梯')
>>> f.close()
# 'rb' 二进制模式直接读取文件的字节串
# 与 utf-8 编码一致
>>> codecs.open(fpath,'rb').read()
'\xe6\xa2\xaf'

>>> fpath='utf8bom.txt'
# encoding='utf-8-sig', 写文件时，会在文件开头加入 字节顺序标记
>>> open(fpath,'w',encoding='utf-8-sig').write('梯')
1
>>> open(fpath,'rb').read()
b'\xef\xbb\xbf\xe6\xa2\xaf'
>>> open(fpath,'r',encoding='utf-8').read()
'\ufeff梯'
# encoding='utf-8-sig', 读文件时，忽略字节顺序标记
>>> open(fpath,'r',encoding='utf-8-sig').read()
'梯'

>>> import sys
# sys.byteorder 当前系统的字节序， 'little'（小端）或 'big'（大端）
>>> sys.byteorder
'little'
# 写文件时，encoding='utf-16' ，默认添加当前系统字节序(little)
>>> open(fpath,'w',encoding='utf-16').write('梯')
1
# 二进制模式，读取实际存放字节串，utf-16小端字节序为 \xff\xfe 
>>> open(fpath,'rb').read()
b'\xff\xfe\xafh'
>>> open(fpath,'r',encoding='utf-16').read()
'梯'
# encoding='utf-16-le'，读文件时，忽略字节顺序标记
>>> open(fpath,'r',encoding='utf-16-le').read()
'\ufeff梯'

# encoding='utf-16-le'，写文件时，不添加BOM
>>> open(fpath,'w',encoding='utf-16-le').write('梯')
1
>>> open(fpath,'rb').read()
b'\xafh'
>>> open(fpath,'r',encoding='utf-16-le').read()
'梯'
# encoding='utf-16-le'，写文件后，不能用 utf-16 读取
>>> open(fpath,'r',encoding='utf-16').read()
Traceback (most recent call last):
  File "<pyshell#102>", line 1, in <module>
    open(fpath,'r',encoding='utf-16').read()
  File "D:\python3\lib\codecs.py", line 322, in decode
    (result, consumed) = self._buffer_decode(data, self.errors, final)
  File "D:\python3\lib\encodings\utf_16.py", line 67, in _buffer_decode
    raise UnicodeError("UTF-16 stream does not start with BOM")
UnicodeError: UTF-16 stream does not start with BOM