Function fails on empty sequence

Question

I am trying to find the date of the newest file in a directory, based on some Python code I shamelessly found on this site.

def newest_file_in_tree(rootfnewer, extension=".avi"):
    return max(
        (os.path.join(dirname, filename)
        for dirname, dirnames, filenames in os.walk(rootfnewer)
        for filename in filenames
        if filename.endswith(extension)),
        key=lambda fn: os.stat(fn).st_mtime)

This works, but fails on empty directories.

    key=lambda fn: os.stat(fn).st_mtime)
ValueError: max() arg is an empty sequence

I concede that I am a Python beginner, and the last line is way beyond my pay grade.

Can anyone suggest how I can trap the error and prevent the function from crashing?

Answer 1

max has a default keyword:

m = max([], key=lambda x: abs(x), default='empty...')
print(m) # 'empty...'

the default will be returned if the iterable is empty. without the default this would raise a ValueError.

---

a different way for you to go would be to use glob in order to list the files:

import os
from glob import glob

def newest_file_in_tree(rootdir, ext='.avi'):
    glob_wildcard = os.path.join(rootdir, '**/*{}'.format(ext))
    glob_gen = glob(glob_wildcard, recursive=True)
    return max(glob_gen, key=lambda fn: os.stat(fn).st_mtime, default='?')

or even more concise with pathlib:

from pathlib import Path

def newest_file_in_tree(rootpath, ext='.avi'):
    return max(Path(rootpath).glob('**/*{}'.format(ext)),
               key=lambda pth: pth.stat().st_mtime, default='?')

Answer 2

You can separate out that monstrously pythonic one liner into two separate lines, to make sure that max is called only when you have a non-empty list of files from which you pick the newest one:

def newest_file_in_tree(rootfnewer, extension=".avi"):
    files = [
        os.path.join(dirname, filename)
        for dirname, dirnames, filenames in os.walk(rootfnewer)
        for filename in filenames
        if filename.endswith(extension)
    ]

    if files:
        return max(files, key=lambda fn: os.stat(fn).st_mtime)

    return None

Not as pretty as @hiroprotagonist's solution, but is more readable.