Any websites you could point me to that describes such a scheme?

(Taking a few chars from name,adr,city to make a unique id and then run 
comparisons rather than trying to match whole columns)

I have 3 databases as it is now. ours with 400.000 records, and two 
external phone lists with over 3M each

thanks

- Frode


> on 8/28/02 10:03 AM, Frode at fstroemm@o... wrote:

> What is a good approach to match customer name,address etc. from my
> company database to the other ones. (To get phone numbers,emails fex.)

Can be a VERY tough process. Depends on the source of the names: Free text
entered by undisciplined operators or imported from purchased name lists.

Say you got stuff from Dun & Bradstreet AND you used their DUNS number. 
That
would give you a very nice unique ID for each entity.

Free text? It will take the comparison of several columns to determine if
you have a dupe or not: Trust the Name the least. Telno, address, zip...

There are schemes where you take the few chars from the name, address, city
and maybe telno and create a unique ID. You might be able to do that
independently on each system and then run your comparisons.

=Alan R.