我可以使用哪个Python库从路径中提取文件名,而不管操作系统或路径格式是什么?

例如,我希望所有这些路径都返回c:

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

当前回答

这是适用于linux和windows以及标准库

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']

def path_leaf(path):
    return path.strip('/').strip('\\').split('/')[-1].split('\\')[-1]

[path_leaf(path) for path in paths]

结果:

['c', 'c', 'c', 'c', 'c', 'c', 'c']

其他回答

也许只是我的all in one解决方案,没有重要的一些新的(考虑tempfile创建临时文件:D)

import tempfile
abc = tempfile.NamedTemporaryFile(dir='/tmp/')
abc.name
abc.name.replace("/", " ").split()[-1] 

获取abc.name的值将是这样的字符串:'/tmp/tmpks5oksk7' 所以我可以用空格.replace("/", " ")替换/,然后调用split()。它会返回一个列表,我得到 列表中最后一个带有[-1]的元素

不需要导入任何模块。

我从来没有见过双开的路,它们存在吗?python模块os的内置特性在这些方面失败了。所有其他工作,还有你用os.path.normpath()给出的警告:

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', 'a/./b/c', 'a\b/c']
for path in paths:
    os.path.basename(os.path.normpath(path))

我在Windows和Ubuntu (WSL)上使用此方法,它只使用“import os”即可工作(我): 基本上,replace()根据当前操作系统平台设置正确的路径分隔符。

如果路径以斜杠'/'结束,那么它不是一个文件而是一个目录,因此它返回一个空字符串。

import os

my_fullpath = r"D:\MY_FOLDER\TEST\20201108\20201108_073751.DNG"
os.path.basename(my_fullpath.replace('\\',os.sep))

my_fullpath = r"/MY_FOLDER/TEST/20201108/20201108_073751.DNG"
os.path.basename(my_fullpath.replace('\\',os.sep))

my_fullpath = r"/MY_FOLDER/TEST/20201108/"
os.path.basename(my_fullpath.replace('\\',os.sep))

my_fullpath = r"/MY_FOLDER/TEST/20201108"
os.path.basename(my_fullpath.replace('\\',os.sep))

在Windows(左)和Ubuntu(通过WSL,右)上:

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。

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。