Is there a best practice or common algorithm for implementing conversion of natural search string for locations location (US only) into its separate components?
for example:
City Name, ST 00000
TO
city => City Name
state => ST
zipcode => 00000
This is for a form so i dont need to handle any possible permutation - i can restrict the format to something like: city, st 00000 but i need to be able to handle omission of the any of the segments within the format, so that they optional to some extent... some examples of supported combinations (case insensitive):
00000 // zipcode
0000-00000 //zipcode
city, st / city and state - comma separated
city st // city and state - space separated
city, st 00000 // city state zip
st 00000 // state and zip - though i only really need the zip
city 00000 // city and zip - though i only really need the zip
I can also use a static set of State abbreviations so those could potentially be matched to validate a state segment if needed.
While i was researching I found some other code referenced in another SO question that i used while i was waiting... I modified the code here to support getting the zipcode as well as the city state: http://www.eotz.com/2008/07/parsing-location-string-php
Others might also find this useful.
@delphist: THANKS. Once i have time to compare accuracy and performance i may switch to your code if its better - its certainly simpler/shorter! If i do Ill mark it as the official answer.