Parsing requires an unambiguous set of rules that detail how
to identify the various parts of an address. If you have a
nicely restricted set of addresses to operate on, it can
probably be done (once you explain the rules), but, in
general, there is no such set of rules for either
people/place names or addresses.
MPD, there is no easy calculation steps involved to parse a mailing address.
There are many industry standard software that do this ($,$$$), but the ones
that do it correctly rely not only on recognition, but on the USPS (Postal
Service) address database. And of course, this is not static and relies on
regular updates. That is talking in general terms. However if your addresses
follow some sort of pattern and are not to crazy or globalized, there may be
an easier approach you can take. Just come back to us with some samples and
we may be able to come up with a handy, but not industry-strength solution.
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.
