> I'm wondering if there's a standard set of routines that we should
> apply to the database to check how smoothly our data migrated.
>
> I've been doing a number of title, author, subject, ISBN, etc.
> searches, for a variety of item formats, checking the appearance of
> the Results & Item displays, etc. But I'm wondering what I might be
> missing by using such a seat-of-the-pants process.
Generally speaking, it's best to identify a bunch of records with
different properties (e.g. serials with a zillion items attached,
something with electronic holdings, stuff in foreign character sets,
summary holdings statements, etc). Also, migrate a test record with
every field you can think of to see if it indexes properly. I don't
know how complex your data was before, but it's a good idea to ask
staff what they're concerned will not make it -- when you do a
migration, "creative" (i.e. nonstandard) use of data points will bite
you in the butt. Don't forget to make sure the right call numbers are
displaying. Most libraries retain all kinds of call numbers that they
don't use, so if you index the wrong one, you've got some serious
problems...
Typically, migrating bib data is a pretty straightforward process and
unless you've misconfigured something, you're probably OK. However,
serials and acq data are far more complex so you'll have to look at
that closely. I've never seen checkin data get migrated smoothly
before, but it's probably happened somewhere. I hope you didn't try to
migrate live circ data since that rarely works out well.
Be aware that fields associated with bib, item, patron, and order
information will not be the same in the old system as the new one.
Some fields will exist in only one system, and some functionality will
simply be different. The important thing to do is make sure that the
data doesn't just move over, but that it loads in a way that you
retain the functionality that you need.
To get back to your original question, I'm not sure there is a
standard series of tests. Library systems analysis is more of a black
art than science ;)
kyle
Received on Wed May 14 2008 - 16:47:11 EDT