Deleteing empty directories

[CinnamonDonkey]

Steven you are right, isDirEmpty() isn't even used. That's what
happens when you try to get a last minute thread going 5 minutes
before home time! ;-)



Thanx for the responses guys! It's been very useful :)




On 30 Mar, 16:38, Steven D'Aprano <st... at REMOVE-THIS-
cybersource.com.au> wrote:
> On Mon, 30 Mar 2009 08:14:55 -0700, CinnamonDonkey wrote:
> > My understanding was that rmtree removes a whole tree not just the empty
> > directories?
>
> So it seems:
>
> >>> os.mkdir('die-die-die')
> >>> os.mkdir('die-die-die/stuff')
> >>> shutil.rmtree('die-die-die')
>
> I think what you want is os.removedirs().
>
> >>> os.makedirs('root/die-die-die/empty/empty/empty')
> >>> os.listdir('root')
>
> ['keep', 'die-die-die']>>> os.removedirs('root/die-die-die/empty/empty/empty')
> >>> os.listdir('root')
>
> ['keep']
>
> > def isDirEmpty( path ):
> >     if not os.path.isdir( path ):
> >         return False
>
> >     contents = os.listdir( path )
>
> >     if len(contents) == 0:
> >         return True
>
> >     return False
>
> That can be simplified to:
>
> # untested
> def isDirEmpty(path):
>     return os.path.isdir(path) and not len(os.listdir(path))
>
> > def RecurseTree( path ):
> >     if not os.path.isdir( path ):
> >         return False
>
> What if it is a symbolic link to a directory?
>
> >     contents = os.listdir( path )
>
> >     if len(contents) == 0:
> >         print "Deleting Empty Dir '%s'" % (path,) #shutil.rmtree(path)
>
> Why do you go to the trouble of defining isDirEmpty() and then not use it?
>
> >     else:
> >         for item in contents:
> >             investigate = "%s\\%s" % (path, item) if
> >             os.path.isdir(investigate):
> >                 RecurseTree( investigate )
>
> As soon as you start recursively walking over directories, you should use
> os.walk. It will almost certainly do what you want.
>
> > if __name__ == '__main__':
> >     RecurseTree( r"c:\temp" )
>
> > But I'm not sure what the max recursion depth is in python?
>
> By default in my version:
>
> >>> sys.getrecursionlimit()
>
> 1000
>
> but it can be changed.
>
> > Plus I think this could be more efficient.
>
> Probably, but why do you care? The file I/O probably will take 99% of the
> time, and I doubt you can improve that.
>
> Of course I could be wrong, so profile, profile, profile, and find out
> where the time really is being spent.
>
> --
> Steven