In Python on a GNU/Linux system, what's the fastest way to recursively scan a directory for all .MOV or .AVI files, and to store them in a list?
Asked
Active
Viewed 1.1k times
8
Keith Pinson
- 7,604
- 6
- 56
- 100
ensnare
- 37,180
- 60
- 149
- 218
-
1Fastest probably involves writing extension to use native code. But do you really want that? – David Heffernan Dec 24 '11 at 17:28
-
Even if you don't want to do that, depending on how many files and directories we're talking about, it might be faster to execute the external `find` command than processing the results of `os.walk()`. But if the `os.walk()` solution is fast enough, it is more elegant and easy to understand/edit. – Michael Hoffman Dec 24 '11 at 17:37
7 Answers
7
You can use os.walk() for recuresive walking and glob.glob() or fnmatch.filter() for file matching:
Check this answer
Community
- 1
- 1
Aleksandra Zalcman
- 3,102
- 1
- 17
- 19
7
I'd use os.walk to scan the directory, os.path.splitext to grab the suffix and filter them myself.
suffixes = set(['.AVI', '.MOV'])
for dirpath, dirnames, filenames in os.walk('.'):
for f in filenames:
if os.path.splitext(f)[1] in suffixes:
yield os.path.join(dirpath, f)
-
This is probably the best solution because it can be easily adapted to enforce case-insensitive matching. – ekhumoro Dec 24 '11 at 19:49
4
Example for a list of files in current directory. You can expand this for specific paths.
import glob
movlist = glob.glob('*.mov')
milancurcic
- 6,072
- 2
- 32
- 46
2
pattern = re.compile('.*\.(mov|MOV|avi|mpg)$')
def fileList(source):
matches = []
for root, dirnames, filenames in os.walk(source):
for filename in filter(lambda name:pattern.match(name),filenames):
matches.append(os.path.join(root, filename))
return matches
Jhonathan
- 1,581
- 1
- 12
- 24
-
The [fnmatch](http://docs.python.org/library/fnmatch.html#module-fnmatch) module only supports very simple glob patterns, so your filter won't work. – ekhumoro Dec 24 '11 at 19:46
-
@ekhumoro if it works, symbols ([],.,?, *, ()) are allowed to glob, python test code and see which works – Jhonathan Dec 24 '11 at 20:01
-
Your pattern is equivalent to `*.[movMOVaipg()]`. This will match, for example, `*.i`, `*.a`, `*.M`, etc, but _not_ `*.MOV`, `*.avi`, etc. Try it for youself! – ekhumoro Dec 24 '11 at 20:21
1
Python 2.x:
import os
def generic_tree_matching(rootdirname, filterfun):
return [
os.path.join(dirname, filename)
for dirname, dirnames, filenames in os.walk(rootdirname)
for filename in filenames
if filterfun(filename)]
def matching_ext(rootdirname, extensions):
"Case sensitive extension matching"
return generic_tree_matching(
rootdirname,
lambda fn: fn.endswith(extensions))
def matching_ext_ci(rootdirname, extensions):
"Case insensitive extension matching"
try:
extensions= extensions.lower()
except AttributeError: # assume it's a sequence of extensions
extensions= tuple(
extension.lower()
for extension in extensions)
return generic_tree_matching(
rootdirname,
lambda fn: fn.lower().endswith(extensions))
Use either matching_ext or matching_ext_ci with arguments the root folder and an extension or a tuple of extensions:
>>> matching_ext(".", (".mov", ".avi"))
tzot
- 87,612
- 28
- 135
- 198
1
I suggest the use of os.walk and a carefully reading of its documentation.
This may be a one liner approach:
[f for root,dirs,files in os.walk('/your/path') for f in files if is_video(f)]
Where in is_video you check your extensions.
Rik Poggi
- 26,862
- 6
- 63
- 81
0
You can also use pathlib for this.
from pathlib import Path
files_mov = list(Path(path).rglob("*.MOV"))
H. Sánchez
- 467
- 5
- 12