Regex for number could contain comma verification#

Even there, I'd be careful - I know that there's a town in England, for example, that contains an exclamation mark ('Westward Ho!') but I have no idea whether there are any street names that contain punctuation marks other than full-stops or apostrophes. '5th Avenue' would be just as legitimate as 'Fifth Avenue' and any regex taking that into account would really be confining itself to disbarring non-alphanumeric characters. A street name, however, really has to be taken as free-text unless you're working against a database of all existing addresses (even that can cause problems with new-builds). A state code is even more friendly in that it's a two-letter code from a nice static list. A zip code, for example, is very IT friendly in that it can be validated in terms of format and I imagine would be verifiable against a list available from the US Mail.

Regex for number could contain comma verification#

If we look at the different parts of an address, we see components with different degrees of potential validation and verification potential.

I'm not overly familiar with US addresses but I'm sure that there must be many places called Whatever Alley/Grove/Square/Crescent/Garden or any number of other things that might never occur to any of us. Addresses are always an absolute pain for both storage and validation so anything that helps with them is good but I suspect that this would exclude an awful lot of perfectly valid addresses.