Thanks for the suggestions.
Option 1 is not possible since I'm guaranteed that there will be only one occurrence of the account number per spreadsheet. So like you say, putting them in a dictionary or other data structure will not buy me much since I still need to go to the iSeries to validate.
Option 2. They are customer accounts that I'm talking about and there are about 700,000+ that I would need to download so that may not work faster.
I may try option 3.
Thanks again for your suggestions...
Thank you,
Antonio Mira
Application Developer - Mid-Ohio Division
Time Warner Cable
1015 Olentangy River Road - 2nd Floor
Columbus, OH 43212
http://www.timewarnercable.com
phone: 614 827 7949
-----Original Message-----
From: systemidotnet-bounces@xxxxxxxxxxxx [mailto:systemidotnet-bounces@xxxxxxxxxxxx] On Behalf Of Walden H. Leverich
Sent: Wednesday, May 13, 2009 2:58 PM
To: .net use with the System i
Subject: Re: [SystemiDotNet] Validating data...
The problem is the number of round trips you're taking, they may be
"fast" but even something that's fast is slow if done enough times.
Three quick options come to mind:
1) Are you caching accounts you've already checked? This is the simplest
change, but might buy you a lot. Easiest thing to do is create a
Dictionary<Account> and then add the ones you lookup, with a "good" or
"no good" as the value. Then check the dictionary before you go to the
400. Obviously this only buys you something if there are duplicate
account numbers.
2) Download the account list into a dictionary and check there. Depends
on the meaning of "account". If we're talking GL accounts and you have
500 of them this makes lots of sense. If we're talking customer accounts
and you have 500,000 of them maybe not so much. But you might be
surprised, still might make sense.
3) Upload the list of accounts to the 400 and do the lookup there.
Instead of doing one by one to them is groups of 10, or 100, or ideally
send all and get back exceptions.
-Walden
midrange.com
