Windows10使用pip安装python包时报错－UnicodeDecodeError:'ascii' codec can't decode byte

本人是Windows10，用的方法2解决的

 

原文链接http://blog.csdn.net/all_over_servlet/article/details/45112221

先交待下开发环境：

  操作系统：Windows 7

  Python版本：2.7.9

  Pip版本：6.1.1

  其他环境忽略

在windows下使用pip下载python包，出现如下错误

 

1.  
   Collecting xxxxxx
2.  
   Exception:
3.  
   Traceback (most recent call last):
4.  
   File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipasecommand.py", line 232, in main
5.  
   status = self.run(options, args)
6.  
   File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipcommandsinstall.py", line 339, in run
7.  
   requirement_set.prepare_files(finder)
8.  
   File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip eq eq_set.py", line 333, in prepare_files
9.  
   upgrade=self.upgrade,
10.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipindex.py", line 305, in find_requirement
11.  
    page = self._get_page(main_index_url, req)
12.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipindex.py", line 783, in _get_page
13.  
    return HTMLPage.get_page(link, req, session=self.session)
14.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipindex.py", line 872, in get_page
15.  
    "Cache-Control": "max-age=600",
16.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestssessions.py", line 473, in get
17.  
    return self.request('GET', url, **kwargs)
18.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipdownload.py", line 365, in request
19.  
    return super(PipSession, self).request(method, url, *args, **kwargs)
20.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestssessions.py", line 461, in request
21.  
    resp = self.send(prep, **send_kwargs)
22.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestssessions.py", line 610, in send
23.  
    r.content
24.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestsmodels.py", line 730, in content
25.  
    self._content = bytes().join(self.iter_content(CONTENT_CHUNK_SIZE)) or bytes()
26.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestsmodels.py", line 655, in generate
27.  
    for chunk in self.raw.stream(chunk_size, decode_content=True):
28.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestspackagesurllib3 esponse.py", line 256, in stream
29.  
    data = self.read(amt=amt, decode_content=decode_content)
30.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendor equestspackagesurllib3 esponse.py", line 186, in read
31.  
    data = self._fp.read(amt)
32.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendorcachecontrolfilewrapper.py", line 54, in read
33.  
    self.__callback(self.__buf.getvalue())
34.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendorcachecontrolcontroller.py", line 217, in cache_response
35.  
    self.serializer.dumps(request, response, body=body),
36.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpipdownload.py", line 268, in set
37.  
    return super(SafeFileCache, self).set(*args, **kwargs)
38.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendorcachecontrolcachesfile_cache.py", line 83, in set
39.  
    with FileLock(name) as lock:
40.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendorlockfilemkdirlockfile.py", line 18, in __init__
41.  
    LockBase.__init__(self, path, threaded, timeout)
42.  
    File "D:Python27libsite-packagespip-6.0.8-py2.7.eggpip\_vendorlockfile\__init__.py", line 189, in __init__
43.  
    hash(self.path)))
44.  
    File "D:Python27lib tpath.py", line 84, in join
45.  
    result_path = result_path + p_path
46.  
    UnicodeDecodeError: 'ascii' codec can't decode byte 0xcb in position 0: ordinal not in range(128)
47.  
     

'UnicodeDecodeError'这个词已经暴露了这个问题是个编码问题

 

什么原因导致了这样的问题？在我的电脑上出现这个问题的原因是由于我的用户目录是中文的，pip在下载的时候调用了这样一行代码

temp_dir = tempfile.mkdtemp('-unpack', 'pip-')

pip把下载的临时文件存放在了用户临时文件中，这个目录一般是C:Users用户名AppDataLocalTemp，目录名中有中文，显然ascii这种编码是不支持的

 

那问题要怎么解决呢？有两种方法解决：

1、把上面的temp_dir那段代码修改一个不包含中文的目录，修改这段代码的文件位置在D:Python27Libsite-packagespip-6.1.1-py2.7.eggpipdownload.py(位置由个人python安装目录决定)

2、修改编码为gbk，修改D:Python27Lib tpath.py(位置由个人python安装目录决定)文件中的def join(path, *paths)函数，在函数内第一行加入

1.  
   # Join two (or more) paths.
2.  
   def join(path, *paths):
3.  
   """Join two or more pathname components, inserting "\" as needed."""
4.  
   reload(sys)
5.  
   sys.setdefaultencoding('gbk')
6.  
   result_drive, result_path = splitdrive(path)
7.  
   for p in paths:
8.  
   p_drive, p_path = splitdrive(p)
9.  
   if p_path and p_path[0] in '\/':
10.  
    # Second path is absolute
11.  
    if p_drive or not result_drive:
12.  
    result_drive = p_drive
13.  
    result_path = p_path
14.  
    continue
15.  
    elif p_drive and p_drive != result_drive:
16.  
    if p_drive.lower() != result_drive.lower():
17.  
    # Different drives => ignore the first path entirely
18.  
    result_drive = p_drive
19.  
    result_path = p_path
20.  
    continue
21.  
    # Same drive in different case
22.  
    result_drive = p_drive
23.  
    # Second path is relative to the first
24.  
    if result_path and result_path[-1] not in '\/':
25.  
    result_path = result_path + '\'
26.  
    result_path = result_path + p_path
27.  
    ## add separator between UNC and non-absolute path
28.  
    if (result_path and result_path[0] not in '\/' and
29.  
    result_drive and result_drive[-1:] != ':'):
30.  
    return result_drive + sep + result_path
31.  
    return result_drive + result_path

注意：

1.  
   reload(sys)
2.  
   sys.setdefaultencoding('gbk')

这两行代码是我后加入的

 

 

一切准备就绪，重新执行pip安装试试吧

 

总结：

1、据说python3的默认编码为'utf-8'，可能不存在这种问题，没有实际测试过

2、这次我直接修改了python和pip中的源码，体现了python是脚本语言的特性

3、如果本文对您有用，请支持原创，谢谢

---------------------------------------------------------------------------------

关注微信公众号即可在手机上查阅，并可接收更多测试分享~