Re: SQL vs native access (was: Record name the same as the filename)

what does faster mean? Is this a step ladder to the moon sort of thing?
 
 
 
---------------------------------------------------------
Booth Martin http://www.MartinVT.com
Booth@xxxxxxxxxxxx
---------------------------------------------------------
 
-------Original Message-------
 
From: RPG programming on the AS400 / iSeries
Date: Thursday, August 21, 2003 3:59:32 PM
To: RPG programming on the AS400 / iSeries
Subject: Re: SQL vs native access (was: Record name the same as the
filename)
 
Rob,
 
Let me be clear: I don't think using SQL is silly (I use it a lot, even
through CLI); I just think using SQL for reading one record (oops, tuple) at
a time is silly.
 
The BPCS situation you describe sounds like an ugly problem to me. Solutions
to ugly problems tend to be ugly themselves, whether they use SQL (how do
you define the host variables?) or native RPG access (of course it can be
done).
 
As for SQL being faster: I know there are situations where SQL is definitely
faster, but I would be mightily surprised if retrieving one record with a
SELECT statement would be faster than an RPG CHAIN. But I must admit I never
tested that.
 
Joep Beckeringh
 
----- Original Message -----
From: <rob@xxxxxxxxx>
To: "RPG programming on the AS400 / iSeries" <rpg400-l@xxxxxxxxxxxx>
Sent: Thursday, August 21, 2003 3:33 PM
Subject: Re: SQL vs native access (was: Record name the same as the
filename)
 
 
> While you may think it is silly there are business cases for this. In our
> case we have a common routines library - ROUTINES. We also will be
> supporting several divisions that use this library. Some of these
> divisions will be using BPCS 405CD, some BPCS 8. In BPCS 8 they changed
> numerous file structures, including KEY sizes! With an SQL based solution
> the same program can support both. With a native rpg access - you cannot.
>
> Another business case. A die hard native person did some time trials and
> was stunned to find the sql faster. However he decided to stick with
> native anyway.
>
> Rob Berendt