Python: get pathname of files using regex pattern recursively

Question

If one want to get the (absolute) file paths (file listings) of all files with names of certain pattern how can this be done in Python (v 3.5 on unix). Something similar to the bash command find -regex 'pattern'. I have been looking as the os, glob, os.path and this SO but cannot get it together.

Say you want absolute paths to the files that matches /.*[pat.txt]$/ and you have the below diretories:

/home/me/dir1/dir1a/filepat.txt #1
/home/me/dir1/dir1a/file.txt
/home/me/dir1/dir1a/filepat.png
/home/me/dir2/filepat.txt #2
/home/me/dir3/dir3a/dir3ab/filepat
/home/me/dir3/dir3a/dir3ac/filepat.txt #3 
/home/me/dir3/dir3a/dir3ac/filepat.png

Then you would get want the three indicated paths:

/home/me/dir1/dir1a/filepat.txt
/home/me/dir2/filepat.txt
/home/me/dir3/dir3a/dir3ac/filepat.txt

One try was:

import fnmatch
import os 
start_path = "/home/me/" 
for root, dirs, files in os.walk(start_path):
    for filename in fnmatch.filter(files, ".*pat.txt"):
        print(os.path.join(start_path, filename))

Python is misspelled in the post title :-) – Apr 20 '17 at 06:26 — , Apr 20 '17 at 06:26
Embarrassing - but good catch – user3375672 Apr 20 '17 at 06:33 — user3375672, Apr 20 '17 at 06:33
don't worry, I was in "debug mode" ;-) – Apr 20 '17 at 07:10 — , Apr 20 '17 at 07:10

score 3 · Answer 1 · answered Apr 20 '17 at 06:30

3

Here is one using regexes, but for the simple case I would go with akash`s answer using in operator

import re
pattern = re.compile(r'.*pat\.txt$')

import fnmatch
import os 
start_path = "/home/me/" 
for root, dirs, files in os.walk(start_path):
    for filename in files:
        if pattern.find(filename):
            print(os.path.join(start_path, filename))

answered Apr 20 '17 at 06:30

redacted

3,789
6
25
38

I was close I see. I think yours is very general though (+1) – user3375672 Apr 20 '17 at 06:36

score 1 · Accepted Answer · answered Apr 20 '17 at 06:27

1

You can use basename and in operator

x = given list
>>> [i for i in x if 'pat.txt' in os.path.basename(i)]
['/home/me/dir1/dir1a/filepat.txt',
 '/home/me/dir2/filepat.txt',
 '/home/me/dir3/dir3a/dir3ac/filepat.txt']

answered Apr 20 '17 at 06:27

akash karothiya

5,736
1
19
29

Oh I completely missed `basename` – user3375672 Apr 20 '17 at 06:36
1

However, this solution will also fetch files like `filepat.txt.bak` – Wiktor Stribiżew Apr 20 '17 at 06:37
To fix that, just use `i.endswith('pat.txt')` – redacted Apr 20 '17 at 06:41
@WiktorStribiżew , we can still achieve this using `endswith('pat.txt')` if required :) – akash karothiya Apr 20 '17 at 06:41
1

- Yes, but then it would fail to fetch `FILEPAT.TXT`. - We may use `lower()`. – Wiktor Stribiżew Apr 20 '17 at 06:46

Python: get pathname of files using regex pattern recursively

2 Answers2