Have you thought about farming it out to a service to do address normalization?
-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of sjl
Sent: Monday, January 09, 2012 12:32 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: Re: SCAN vs %scan
I have around 750,000 records in one file which have fields Name, Address1,
Address2, City, and County and almost 500,000 records in another file which
have Fullname, Lastname, Firstname, Middlename, DBA Name, Address1,
Address2, City, and County.
Each of the Name fields must be run through this routine.
Each of the Address-related fields must be run through a similar routine.
- sjl
"Luis Rodriguez" wrote in message
news:mailman.3251.1326140553.2619.rpg400-l@xxxxxxxxxxxx...
OK., just for the heck of it I ran some (totally) unscientific tests:
I wrote a small program that loaded a 4000 element array from an old
customer file, and then tried to find the same string in two subroutines:
The first using %SCAN and the second using SCAN. Ran the program several
times with and without optimization and the result was that %SCAN seemed to
be consistently faster than SCAN.
This was run on an old V5R3 Power 5, and the results varied a lot (it
depends on what was being done on the server at the moment), but the time
was always at least a second faster using %SCAN.
Your mileage can (and sure will) vary :-)
HTH,
Luis Rodriguez
IBM Certified Systems Expert - eServer i5 iSeries
midrange.com
