Stephen,
I have done something similar where I work. The difficult part is
putting the multiple databases together into one. What I did is I took
one database as my starting point. Then I took a second database and
designed a routine that looped though each record of the second
database to see if that individual was already in my first db. I check
for the person's first name and last name. If it was not found, I then
check to see if there were any records in the first db that had a
variation of the person's first name and matching last name. For
example, Bill and William, John and Jonathan, Sue and Susan, etc. If a
duplicate was found, I saved a record to a table called Duplicates that
contained the data from the record I was checking as well as the
possible duplicate from the database. I then went back and checked the
possible duplicates. If I found a match, I deleted it. If the
possible duplicate was not a match, I then inserted it into the
database.

I hope that gives you a little direction at least.

Jeff L wrote:
I've done something similar for a customer DB on a project a few years
ago. A couple things to consider:
* validating the addresses
* matching name AND address.

There are some software products you can buy (not cheap) that validate
your addresses against the US Postal Service address database. They
used to have a place on their website (usps.com) where you could
manually validate an address against thier DB. Looks like you need
special permission to use that service now. You'll find the link for it
under the Zipcode look up.

When matching names and addresses, it is sometimes easier to match last
first. If you have a potential duplicate based on name, then try
matching on the address. It might be that Bill Jones is the son of
William Jones. And while they live in the same city, they on different
parts of town (different street addresses) Or Bill Jones moved when he
was promoted to VP and now lists his name more formally to go along
with his new job.

For a club you might make a report of potential dups and get some human
input into what is the real situation.

HTH,
Ed

the data structure, de-dup the records from all databases,
load into uniform structure.