13.1. File Path

  • Relative paths are relative to the current working directory

  • Absolute paths start with root directory

  • Python works with both relative and absolute path

13.1.1. Relative Paths

  • Path is relative to currently running script

  • Relative paths works the same on Windows and *nix (Linux, macOS, BSD, etc.)

  • . - Current directory

  • .. - Parent directory

Current directory:

>>> file = 'myfile.csv'
>>> file = 'data/myfile.csv'
>>> file = 'data/csv/myfile.csv'

>>> file = './myfile.csv'
>>> file = './data/myfile.csv'
>>> file = './data/csv/myfile.csv'

Parent directory:

>>> file = '../myfile.csv'
>>> file = '../data/myfile.csv'
>>> file = '../data/csv/myfile.csv'

>>> file = '../../myfile.csv'
>>> file = '../../data/myfile.csv'
>>> file = '../../data/csv/myfile.csv'

>>> file = '../../../myfile.csv'
>>> file = '../../../data/myfile.csv'
>>> file = '../../../data/csv/myfile.csv'

13.1.2. Absolute Paths

  • Absolute path include all entries in the directories hierarchy

  • Absolute path on *nix starts with root / dir

  • Absolute path on Windows starts with drive letter

Linux (and other *nix):

>>> file = '/home/mwatney/myfile.csv'

macOS:

>>> file = '/Users/mwatney/myfile.csv'

Windows:

>>> file = 'c:/Users/mwatney/myfile.csv'

13.1.3. Windows Path Problem

  • Problem with paths on Windows

  • Use backslash (\\) as a path separator

  • Use r-string for paths

Let's say we have a path to a file:

>>> print('C:/Users/mwatney/newfile.txt')
C:/Users/mwatney/newfile.txt

Paths on Windows do not use slashes (/). You must use backslash (\\) as a path separator. This is where all problems starts. Let's start changing slashes to backslashes from the end (the one before newfile.txt):

>>> print('C:/Users/mwatney\newfile.txt')
C:/Users/mwatney
ewfile.txt

This is because \n is a newline character. In order this to work we need to escape it.

Now lets convert another slash to backslash, this time the one before directory named mwatney:

>>> print('C:/Users\mwatney\newfile.txt')  
SyntaxWarning: invalid escape sequence '\m'
C:/Users/mwatney
ewfile.txt

Since Python 3.12 all non-existing escape characters (in this case \m will need to be escaped or put inside of a row strings. This is only a warning (SyntaxWarning: invalid escape sequence '\m', so we can ignore it, but this behavior will be default sometime in the future, so it is better to avoid it now.

The last slash (the one before Users):

>>> print('C:\Users\\mwatney\\newfile.txt')
Traceback (most recent call last):
SyntaxError: (unicode error) 'unicodeescape' codec can't decode bytes in position 2-3: truncated \UXXXXXXXX escape

This time the problem is more serious. Problem is with \Users. After escape sequence \U Python expects hexadecimal Unicode codepoint, i.e. \U0001F600 which is a smiley 😀 emoticon emoticon. In this example, Python finds letter s, which is invalid hexadecimal character and therefore raises an SyntaxError telling user that there is an error with decoding bytes. The only valid hexadecimal numbers are 0123456789abcdefABCDEF and letter s isn't one of them.

There is two ways how you can avoid this problem. Using escape before every slash:

>>> print('C:\\Users\\mwatney\\newfile.txt')
C:\Users\mwatney\newfile.txt

Or use r-string:

>>> print(r'C:\Users\mwatney\newfile.txt')
C:\Users\mwatney\newfile.txt

Both will generate the same output, so you can choose either one. In my opinion r-strings are less error prone and I use them each time when I have to deal with paths.

13.1.4. Convention

  • Never hardcode paths, use constant as a file name or file path

  • Convention (singular form): FILE, FILENAME, FILEPATH, PATH

  • Convention (plural form): FILES, FILENAMES, FILEPATHS, PATHS

  • Note, that PATH is usually used for other purposes (sys.path or os.getenv('PATH'))

>>> FILE = 'myfile.txt'

>>> FILES = [
...     'myfile1.txt',
...     'myfile2.txt',
...     'myfile3.txt',
... ]

13.1.5. Recap

  • Relative paths are relative to the current working directory

  • Absolute paths start with root directory

  • Python works with both relative and absolute path

  • Path separator / (slash) is used on *nix operating systems: Linux, macOS, BSD and other POSIX compliant OSes (excluding older versions of Windows)

  • Path separator \ (backslash) is used on Windows

  • In newer Windows versions both \ and / works the same

  • . - Current directory

  • .. - Parent directory

  • Absolute path on *nix starts with root / dir

  • Absolute path on Windows starts with drive letter

  • Never hardcode paths, use constant as a file name or file path

  • On Windows use r-string for paths