script to "best match" text by comparing to table

[carlitos]

Hi everyone. I'm new to this forum. I've been searching for a script but have been unseccessful until now:queasy:.

My problem is the following (this is not USA but Spain):

I have millions of addresses that include a district field. Many of those districts are badly written. I also have a post code field for each address.

On the other hand, I have a table of all postcodes with all possible districts

A postcode can have different districts.

A district can have different postcodes.

I need to clean the wrongly written districts using the correctly written one from the postcode table.

Do you know of a script that can look for "best match" to the wrongly written district and replace the wrong value with the correct one?

Thanks!

[dagon]

It wont be easy but have a look at

similar_text()
levenshtein()
soundex()
metaphone()

all in the manual

[ednark]

there are also some services like yahoo, google, usps, and ups that offer free addressing services (us only for the most part). you would send off the address, and they would correct it for you. try googling around for them and their apis.