Hi everyone,
We're looking at doing some broad clean-up of all of our bad address records - getting all the bad zip codes to be the same, with the same city and state, among other things. Anyone else done this and have any pieces of advice, technically speaking or otherwise? I'm nervous to operate on this many records in our live environment but would prefer specific fear to general!
Thanks!
Frannie
Hi Frannie,
We do a lot of that at PRAC, but I agree with Beth that testing is essential. We also work with our users to make sure that they're following our consortium's data standards to minimize the new bad addresses we see. For example, we ask our organizations to only use their mailing address when it is valid for a patron and instead use a dummy address (street1=Unknown, zip=99999) for patrons when they don't give an address. (Our consortium requires addresses, but that's another discussion.)
With that in mind, I've taken a slightly different approach to "correcting" bad addresses, which is to have a view that lists common bad address strings. Then, rather than changing the address itself, I use that view to apply an "Invalid Address" address type. That way we keep any address information that might have helped identify the patron.
If you're looking for ways to keep an eye on this issue going forward as well, you might want to check out the shared reports section in TASK, filtering by "address". There are a number of address-related utilities and reports, including one that I shared a few years ago that points out common US address issues.
Good luck!
~Katie