RE: "Reference #"

 
If the file is created with auto_increment then that is enforced regardless
of how records are added/inserted to the file?
 
 
---------------------------------------------------------
Booth Martin   http://www.MartinVT.com
Booth@xxxxxxxxxxxx
---------------------------------------------------------
 
-------Original Message-------
 
From: RPG programming on the AS400 / iSeries
Date: Friday, March 28, 2003 17:40:59
To: 'RPG programming on the AS400 / iSeries'
Subject: RE: "Reference #"
 
It would appear (to me) that the SQL auto-increment soulution is by far the
best solution. Management of the field, and its values, is handled by the
DBMS with less potential for error. Performance will far exceed that of a
trigger-based solution.
 
Eric DeLong
Sally Beauty Company
MIS-Project Manager (BSG)
940-898-7863 or ext. 1863
 
 
 
-----Original Message-----
From: Booth Martin [mailto:Booth@xxxxxxxxxxxx]
Sent: Friday, March 28, 2003 2:32 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: RE: "Reference #"
 
 
mainly because if you use a sequential number you have to maintain it, back
it up, restore it, view it, etc. Why bother?
 
---------------------------------------------------------
Booth Martin http://www.MartinVT.com
Booth@xxxxxxxxxxxx
---------------------------------------------------------
 
-------Original Message-------
 
From: RPG programming on the AS400 / iSeries
Date: Friday, March 28, 2003 14:46:44
To: RPG programming on the AS400 / iSeries
Subject: RE: "Reference #"
 
What's the advantage of using a random number? It seems to me that a
sequential number will always be unique, and will certainly be faster to
generate.
 
Why the random number? Are you trying to prevent the user from being
able to guess what the number will be?
 
I agree with you about timestamps... they are not a good idea.
 
On Fri, 28 Mar 2003, Booth Martin wrote:
 
> There's no purpose served by using a sequential number except uniqueness.
> If you want uniqueness just go for a random number. The chances of a
random
> number duplicating are slim, and the delay to renumber in those few
> instances is a small delay indeed. Forget about using the timestamp for
> uniqueness. Its troublesone and buys you nothing.