Use one glob.glob instead of multiple glob.glob

Question

I have a folder containing files in different types. I only need to deal with image files (jpg, png, jpeg, JPG, ...). For that, I use glob.glob to get only files with 'jpg', 'png', ... extensions.
The problem is that I have to use glob.glob many times (as much as the image extensions number) to get all the images:

import glob
images = glob.glob('tests/*.jpg') + glob.glob('tests/*.jpeg') + glob.glob('tests/*.png') + glob.glob('tests/*.JPG')

Is there a way to use glob.glob in a more optimized form (something like glob.glob('path/*.{'jpg', 'png', 'jpeg', 'JPG'}) or is there a simpler function to use it instead of glob?

`glob.glob('tests/*.{jpg, png, JPG}')` returns an empty list — singrium, Feb 12 '19 at 15:49
Sorry, looks like it only support simple shell-style wildcards — han solo, Feb 12 '19 at 15:52
Although you could do something like `[x for x in glob.glob('*') if os.path.splitext(x)[1] in ['.txt', '.jpg', '.png'] ]` :) — han solo, Feb 12 '19 at 15:55
IIRC glob is case-insensitive, so at the very least you can remove the check for "JPG", since "jpg" will cover that. — Kevin, Feb 12 '19 at 15:58
@brunodesthuilliers, oh, maybe it's OS-specific. I'm on Windows and `glob("*.txt")` returns the same results as `glob("*.TXT")`. — Kevin, Feb 12 '19 at 16:10

Alex · Answer 1 · 2019-02-12T16:23:17.363

4

Using pathlib:

from pathlib import Path

extensions = ['.jpg', '.png', '.jpeg']
images = [x for x in Path('tests').iterdir() if x.suffix.lower() in extensions]

edited Feb 12 '19 at 16:23

answered Feb 12 '19 at 15:52

Alex

6,610
3
20
38

When I run the code you suggested, I got this output: `[PosixPath('tests/a.jpg'), PosixPath('tests/b.jpg'), PosixPath('tests/d.png'), PosixPath('tests/e.jpeg'), PosixPath('tests/c.JPG')] `, what `PosixPath()` stands for? – singrium Feb 12 '19 at 15:57
1

@singrium what about reading the FineManual ? https://docs.python.org/3/library/pathlib.html – bruno desthuilliers Feb 12 '19 at 15:58
1

These are described here: https://docs.python.org/3/library/pathlib.html#pure-paths. They are a platform independent way of representing files. – Alex Feb 12 '19 at 15:59
1

@brunodesthuilliers, I'll take a look at the documentation. Thank you. – singrium Feb 12 '19 at 16:01

score 3 · Accepted Answer · answered Feb 12 '19 at 16:02

You cannot do complex wildcard globbing like glob.glob('*.{JPG, png..}) because, if you take a look at the source code you will see,

def glob(pathname):
    """Return a list of paths matching a pathname pattern.
    ....
    """
    return list(iglob(pathname))

And then if you find the source of iglob you will then see,

def iglob(pathname):
    ....
    ....
    dirname, basename = os.path.split(pathname)

    # voila, here, our complex glob wildcard will certainly break, and can't be used :)

Therefore, you can only do simple globbing using glob :)

Use one glob.glob instead of multiple glob.glob

2 Answers2