Question

Looking for a python regex pattern. Seems like it has to exist, but it has me stumped.

If I need to find an address, and the strings I am searching can be of the form

address_is_after_123
 - or -
123_address_is_before

Note, there could be more than two permutations, but I’m hoping a solution for two permutations could be extended to more.

I could simply create multiple regexes, but I’d ideally like a single regex. The best I’ve got is:

m = re.match("(?:address_is_after_(\d+)|(\d+)_address_is_before)",text)

This works, but the I have to test whether m.group(1) or m.group(2) has the value. Is there a way to write the regex so that if it matches I can grab the address without additional processing?

Answer 1

You could do it with lookarounds, provided that the length of the lookbehind ("address_is_after_") is constant:

>>> m = re.search(r"(?<=address_is_after_)\d+|\d+(?=_address_is_before)",text)
>>> m.group(0)
'123'