这是一个仅适用于regex的解决方案，它似乎适用于任何OS上的任何OS路径。

不需要其他模块，也不需要预处理:

import re

def extract_basename(path):
  """Extracts basename of a given path. Should Work with any OS Path on any OS"""
  basename = re.search(r'[^\\/]+(?=[\\/]?$)', path)
  if basename:
    return basename.group(0)


paths = ['a/b/c/', 'a/b/c', '\\a\\b\\c', '\\a\\b\\c\\', 'a\\b\\c',
         'a/b/../../a/b/c/', 'a/b/../../a/b/c']

print([extract_basename(path) for path in paths])
# ['c', 'c', 'c', 'c', 'c', 'c', 'c']


extra_paths = ['C:\\', 'alone', '/a/space in filename', 'C:\\multi\nline']

print([extract_basename(path) for path in extra_paths])
# ['C:', 'alone', 'space in filename', 'multi\nline']

更新:

If you only want a potential filename, if present (i.e., /a/b/ is a dir and so is c:\windows\), change the regex to: r'[^\\/]+(?![\\/])$' . For the "regex challenged," this changes the positive forward lookahead for some sort of slash to a negative forward lookahead, causing pathnames that end with said slash to return nothing instead of the last sub-directory in the pathname. Of course there is no guarantee that the potential filename actually refers to a file and for that os.path.is_dir() or os.path.is_file() would need to be employed.

这将匹配如下:

/a/b/c/             # nothing, pathname ends with the dir 'c'
c:\windows\         # nothing, pathname ends with the dir 'windows'
c:hello.txt         # matches potential filename 'hello.txt'
~it_s_me/.bashrc    # matches potential filename '.bashrc'
c:\windows\system32 # matches potential filename 'system32', except
                    # that is obviously a dir. os.path.is_dir()
                    # should be used to tell us for sure

正则表达式可以在这里测试。

在Python 2和3中，使用模块pathlib2:

import posixpath  # to generate unix paths
from pathlib2 import PurePath, PureWindowsPath, PurePosixPath

def path2unix(path, nojoin=True, fromwinpath=False):
    """From a path given in any format, converts to posix path format
    fromwinpath=True forces the input path to be recognized as a Windows path (useful on Unix machines to unit test Windows paths)"""
    if not path:
        return path
    if fromwinpath:
        pathparts = list(PureWindowsPath(path).parts)
    else:
        pathparts = list(PurePath(path).parts)
    if nojoin:
        return pathparts
    else:
        return posixpath.join(*pathparts)

用法:

In [9]: path2unix('lala/lolo/haha.dat')
Out[9]: ['lala', 'lolo', 'haha.dat']

In [10]: path2unix(r'C:\lala/lolo/haha.dat')
Out[10]: ['C:\\', 'lala', 'lolo', 'haha.dat']

In [11]: path2unix(r'C:\lala/lolo/haha.dat') # works even with malformatted cases mixing both Windows and Linux path separators
Out[11]: ['C:\\', 'lala', 'lolo', 'haha.dat']

使用您的测试用例:

In [12]: testcase = paths = ['a/b/c/', 'a/b/c', '\\a\\b\\c', '\\a\\b\\c\\', 'a\\b\\c',
    ...: ...     'a/b/../../a/b/c/', 'a/b/../../a/b/c']

In [14]: for t in testcase:
    ...:     print(path2unix(t)[-1])
    ...:
    ...:
c
c
c
c
c
c
c

这里的思想是将所有路径转换为pathlib2的统一内部表示形式，根据平台使用不同的解码器。幸运的是，pathlib2包含一个名为PurePath的通用解码器，它可以在任何路径上工作。如果这不起作用，您可以使用fromwinpath=True强制识别windows路径。这将把输入字符串分成几个部分，最后一个是你要找的叶子，因此是path2unix(t)[-1]。

如果参数nojoin=False，则路径将被连接回来，因此输出只是转换为Unix格式的输入字符串，这对于跨平台比较子路径非常有用。

像其他人建议的那样使用os.path.split或os.path.basename并不能在所有情况下工作:如果您在Linux上运行脚本并试图处理经典的windows样式的路径，它将失败。

Windows路径可以使用反斜杠或正斜杠作为路径分隔符。因此，ntpath模块(相当于os. path)在windows上运行时的路径)将适用于所有平台上的所有(1)路径。

import ntpath
ntpath.basename("a/b/c")

当然，如果文件以斜杠结束，basename将为空，所以创建自己的函数来处理它:

def path_leaf(path):
    head, tail = ntpath.split(path)
    return tail or ntpath.basename(head)

验证:

>>> paths = ['a/b/c/', 'a/b/c', '\\a\\b\\c', '\\a\\b\\c\\', 'a\\b\\c', 
...     'a/b/../../a/b/c/', 'a/b/../../a/b/c']
>>> [path_leaf(path) for path in paths]
['c', 'c', 'c', 'c', 'c', 'c', 'c']

(1) There's one caveat: Linux filenames may contain backslashes. So on linux, r'a/b\c' always refers to the file b\c in the a folder, while on Windows, it always refers to the c file in the b subfolder of the a folder. So when both forward and backward slashes are used in a path, you need to know the associated platform to be able to interpret it correctly. In practice it's usually safe to assume it's a windows path since backslashes are seldom used in Linux filenames, but keep this in mind when you code so you don't create accidental security holes.

如果您的文件路径不是以“/”结尾，且目录以“/”分隔，则使用以下代码。众所周知，path通常不以“/”结尾。

import os
path_str = "/var/www/index.html"
print(os.path.basename(path_str))

但在某些情况下，像url以“/”结尾，然后使用以下代码

import os
path_str = "/home/some_str/last_str/"
split_path = path_str.rsplit("/",1)
print(os.path.basename(split_path[0]))

但是当你的路径被“\”分开时，你通常在Windows路径中找到，然后你可以使用以下代码

import os
path_str = "c:\\var\www\index.html"
print(os.path.basename(path_str))

import os
path_str = "c:\\home\some_str\last_str\\"
split_path = path_str.rsplit("\\",1)
print(os.path.basename(split_path[0]))

您可以通过检查操作系统类型将两者组合成一个函数并返回结果。

为了完整起见，这里是python 3.2+的pathlib解决方案:

>>> from pathlib import PureWindowsPath

>>> paths = ['a/b/c/', 'a/b/c', '\\a\\b\\c', '\\a\\b\\c\\', 'a\\b\\c', 
...          'a/b/../../a/b/c/', 'a/b/../../a/b/c']

>>> [PureWindowsPath(path).name for path in paths]
['c', 'c', 'c', 'c', 'c', 'c', 'c']

这在Windows和Linux上都适用。

这是工作!

os.path.basename(name)

但是你不能在Linux中通过Windows文件路径获取文件名。Windows。 操作系统。不同操作系统上不同模块的路径加载:

Linux - posixpath Windows - npath

所以你可以用os。路径总是得到正确的结果