On Aug 3, 11:44 am, Robert Kern <robert.k...@gmail.com> wrote: > On 2009-08-03 12:29, MRAB wrote: > > > > > > > Robert Kern wrote: > > [snip] > > >> for line in readThis: > >> key_match = key.search(line) > >> if key_match is not None: > >> this_key = key_match.group(1) > >> # ... do something with this_key > >> map_match = map.search(line) > >> if map_match is not None: > >> this_map = map_match.group(1) > >> # ... do something with this_map > >> parcel_match = parcel.search(line) > >> if parcel_match is not None: > >> this_parcel = parcel_match.group(1) > >> # ... do something with this_parcel > > > re.search and re.match will return a MatchObject if successful or None > > if unsuccessful. A MatchObject is always true, so you can simplify to: > > > ... > > if key_match: > > this_key = key_match.group(1) > > # ... do something with this_key > > ... > > True, but I prefer to maintain a consistent style for None-testing regardless > of > whether or not I am sure the not-None type will have a suitable __nonzero__ > implementation.
There's one thing to keep in mind. In the unknown future, it's possible that re.match could be changed to return an object with a false value that contains some information about the reason for failure, or partial match information, or something like that. If that happens, code like: if m: will continue to work, while code like if m is not None: will suddenly fail. Granted, it's not likely to happen for regexps but it could happen in similar situations in other places. I am one of the biggest critics of implicit boolean values, but given that Python has it, I think it is best to adhere strongly idiomatic uses like regexps. Carl Banks -- http://mail.python.org/mailman/listinfo/python-list
