How to use glob to read limited set of files with

问题:

How to use glob to only read limited set of files?

I have json files named numbers from 50 to 20000 (e.g. 50.json,51.json,52.json...19999.json,20000.json) within the same directory. I want to read only the files numbered from 15000 to 18000.

To do so I'm using a glob, as shown below, but it generates an empty list every time I try to filter out for the numbers. I've tried my best to follow this link (https://docs.python.org/2/library/glob.html), but I'm not sure what I'm doing wrong.

>>> directory = "/Users/Chris/Dropbox"
>>> read_files = glob.glob(directory+"/[15000-18000].*")
>>> print read_files
[]

Also, what if I wanted files with any number greater than 18000?

回答1:

You are using the glob syntax incorrectly; the [..] sequence works per character. The following glob would match your files correctly instead:

'1[5-8][0-9][0-9][0-9].*'

Under the covers, glob uses fnmatch which translates the pattern to a regular expression. Your pattern translates to:

>>> import fnmatch
>>> fnmatch.translate('[15000-18000].*')
'[15000-18000]\\..*\\Z(?ms)'

which matches 1 character before the ., a 0, 1, 5 or 8. Nothing else.

glob patterns are quite limited; matching numeric ranges is not easy with it; you'd have to create separate globs for ranges, for example (glob('1[8-9][0-9][0-9][0-9]') + glob('2[0-9][0-9][0-9][0-9]'), etc.).

Do your own filtering instead:

directory = "/Users/Chris/Dropbox"

for filename in os.listdir(directory):
    basename, ext = os.path.splitext(filename)
    if ext != '.json':
        continue
    try:
        number = int(basename)
    except ValueError:
        continue  # not numeric
    if 18000 <= number <= 19000:
        # process file
        filename = os.path.join(directory, filename)

回答2:

Although it hardly counts as beautiful code, you could implement your own filtering as follows:

import os, re
directory = "/Users/Chris/Dropbox"
all_files = os.listdir(directory)

read_files = [this_file for this_file in all_files 
                if (int(re.findall('\d+', this_file)[-1]) > 18000)]

print read_files

The crucial line here (should) iterate through each file name in the directory (for this_file in all_files), pull out a list of number segments in that file name (re.findall('\d+', this_file)), and include it in read_files if the last of these number segments, as an integer, is greater than 18000.

I think this will break on files with no integers in the name, so user beware.

---

Edit: I see the previous answer has been edited to include what looks a much better thought out way to do this.