Question

I’m trying to search for a string that has 6 digits, but no more, other chars may follow. This is the regex I use \d{6}[^\d] For some reason it doesn’t catch the digits which \d{6} do catch.

Update

Now I’m using the regex (\d{6}\D*)$ which do makes sence. But I can’t get it to work anyways.

Update 2 – solution

I should of course grouped the \d{6} with parentheses. Doh! Otherwise it includes the none-digit and tries to make a date with that.

End of update

What I’m trying to achive (as a rather dirty hack) is to find a datestring in the header of a openoffice document in either of the following formats: YYMMDD, YYYY-MM-DD or YYYYMMDD. If it finds one of these (and only one) it set the mtime and atime of that file to that date. Try to create a odt-file in /tmp with 100101 in the header and run this script (sample file to download: http://db.tt/9aBaIqqa). It should’nt according to my tests change the mtime/atime. But it will change them if you remove the \D in the script below.

This is all of my source:

import zipfile
import re
import glob
import time
import os

class OdfExtractor:
    def __init__(self,filename):
        """
        Open an ODF file.
        """
        self._odf = zipfile.ZipFile(filename)

    def getcontent(self): 
        # Read file with header
        return self._odf.read('styles.xml')

if __name__ == '__main__':
    filepattern = '/tmp/*.odt'

    # Possible date formats I've used
    patterns = [('\d{6}\D', '%y%m%d'), ('\d{4}-\d\d-\d\d', '%Y-%m-%d'), ('\d{8}', '%Y%m%d')]

    # go thru all those files
    for f in glob.glob(filepattern):
        # Extract data
        odf = OdfExtractor(f)

        # Create a list for all dates that will be found
        findings = []

        # Try finding date matches
        contents = odf.getcontent()
        for p in patterns:
            matches = re.findall(p[0], contents)
            for m in matches:
                try:
                    # Collect regexp matches that really are dates
                    findings.append(time.strptime(m, p[1]))
                except ValueError:
                    pass

        print f
        if len(findings) == 1: # Don't change if multiple dates was found in file
            print 'ändrar till:', findings[0]
            newtime = time.mktime(findings[0])
            os.utime(f, (newtime, newtime))
        print '-' * 8

Answer 1

I was pretty stupid. If I add an \D to the end of the search the search will of course return that none digit also which I did’nt want. I had to add parenthesis to the part I really wanted. I feel pretty stupid for not catching this with a simple print statement after loop. I really need to code more frequently.