I know this is a very old message.. and I see you got lots of good answers. My question is what is the "fuzzy" matching technique.
I'm not so concerned about interface. I am looking for a good VFP based matching algorithm. e.g. something that recognizes spelling mistakes, maybe strip's out commas, abbreviations, identifies singular vs plural, use a combination of = and like operator, etc.

>Hi,
>I have a system where customers get manually added regularly to a table which results in duplicate customers. I need to try and merge duplicates into one record (by merge I am just taking the last record entered or letting the user decide which one to keep and then getting rid of the other records after assigning all the customer's transactions to the record being kept). I can check for duplicates based on customer name, address, ID Card#, Phone# and I do so using a "fuzzy" matching technique.
>
>I am trying to create an interface that will make this work well and am getting stuck, so I'd appreciate any ideas.
>
>I can let the user pick one customer from the list of customers and then search for duplicates, present the duplicates in a multi select grid and let the user select the duplicate records to merge into the originally selected customer. This has the drawback of the user needing to know that a customer record has duplicates, so it won't work well in the situation where the user wants a list of all records that have potential duplicates.
>
>So I need to let the user do a search on all customer records and present possible duplicates for each record, then select the record to keep and then select the records to "merge" into the one to keep, bearing in mind that I may have any number of different customers that I am trying to merge at the same time.
>
>Is that confusing or what!? :)