Wikipedia:Reference desk/Archives/Computing/2020 July 16

Welcome to the Wikipedia Computing Reference Desk Archives
The page you are currently viewing is a transcluded archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

July 16

does the file exist?

In Python, how can I ask whether a file exists that matches a partial name (e.g. archive*.zip)? —Tamfang (talk) 00:13, 16 July 2020 (UTC)

Use glob.glob -- Finlay McWalter··–·Talk 11:33, 16 July 2020 (UTC)

Strictly, glob.glob is eager, which might be an issue if you have a potentially very large number of matches, and/or a slow folder (e.g. on a network share) - because it will read all the directory entries, even if you only care about whether there is a single match. The solution, iff you actually have this case, is to use glob.iglob, which is a generator, and thus is lazy. e.g.:

import glob

def exists(f):
    try:
        next(glob.iglob(f))
        return True
    except:    
        return False

-- Finlay McWalter··–·Talk 16:00, 16 July 2020 (UTC)

However, I'll add that checking for existence directly is only needed if you want to do something such as spit out information to the user, like "directory foo contains 0 files matching archive*.zip". If you want to operate on the files, just try doing so and catch the resulting exception if there are no matches. (FileNotFoundException I think? Not a Python guru.) This is one of the things that's handy about exceptions. No checking for fifteen different things like in C and I hope you didn't forget any, like checking whether a pointer is null! Oops, you just introduced a security hole that someone will come along and exploit later! --47.146.63.87 (talk) 18:45, 16 July 2020 (UTC)

Thanks y'all! Here's my story. I generate lots of algorithmic image files, and my convention is to name them [parameters]-[creator].png, thus a file made by foo.py will have a name ending in *-foo.png. I want to skip generating the file if it has already been made by any program, not only the same program, hence a check for [parameters]-*.png. (One of the "parameters" is usually generated by the program itself.) —Tamfang (talk) 00:48, 17 July 2020 (UTC)

Right, you can use the x mode with open, and it will fail (throwing an exception) if the file already exists. --47.146.63.87 (talk) 01:42, 17 July 2020 (UTC)

Oops, never mind. Read that wrong the first time. --47.146.63.87 (talk) 06:13, 17 July 2020 (UTC)

any(glob.iglob(f)) should get rid of the need for dealing with the exception. 2601:648:8202:96B0:0:0:0:5B74 (talk) 20:21, 23 July 2020 (UTC)