Re: HELP! Looking for Duplicate Record Eliminator!

References: <446teu$plt@sundog.tiac.net> Distribution: Steve Berube (sberube@tiac.net) wrote: : Hi, sorry to post in these groups, but I didn't know where to ask :( : I need a windows/dos based program which can remove duplicate records from : multiple databases at a time : IE DATABASE 1 Contains Company Name XYZ Inc. And Database 2 Contains Company Name : XYZ Incorporated. I need a semi-intelligent program to say ok these match, : which one do you want to keep?? Any help appreciated!!! : Mail me at SBERUBE@TIAC.NET : Thanks : - Steve Sorry, I do not know of any packages set-up for this (I am assuming Informix databases. However, I do have a general suggestion if you have someone that can do the programming. The biggest pain for deduplication of names/addresses is the variations of entering the street addresses. For my databases with address, I keep a "parsed" or "fielded" street address along with the original *raw* inputted street. This not only allows you to do deduplication, but is great for sorting by street address if any kind of routing is necessary. The columns used for the "fielded" street address: House Number INTEGER Half Address? CHAR(01) (Y if house number includes: 1/2) Directional CHAR(02) (N,S,E,W,NE,NW,SE,SW) Street Name CHAR(20) Street Type CHAR(02) (ST,RD,AV,BL,PW,.....) Unit Number CHAR(05) P. O. Box CHAR(10) Unit Numbers contain Apt numbers, office #s, Building #s or even Rural Route #s. The Street Type abbreviation standards are available from the Postal Svc. What ever is left over after searching in the logic for the other fields is put into the street name. If creating the fielded street online from operator then you could also build a Street Name table (especially for addresses geographically clustered in a small area) where the operator is warned if the street name entered is not on file and asked to verify/add. For Example, if three operators enter one street address in these formats: 1. 2 North West Main Street Apt 105 2. 2 N. W. Main Str # 105 3. Two N W Main Str #105 They would all be stored in the record (uppercased) as: House Number: 2 Half? N Directional NW Street Name MAIN Street Type ST Unit Number 105 Good luck in your search. And let ME know if you find anything. Joe Nay chpr@usa.net