Re: Different result from full outer join in ACS RunSqlScript and SQLRPGLE

[VERNON HAMBERG Owner <vhamberg@xxxxxxxxxxxxxxx>]

Hi Charles


Check my post, it has PTFs for 7.3, 7.4, and 7.5 - Martijn had provided the 7.4 there, as well as the APAR.


RegardsVern


On Wed, 22 Nov, 2023 at 8:39 AM, Charles Wilt <charles.wilt@xxxxxxxxx> wrote:
 

To: rpg programming on ibm i

I did say it was possible, and to contact IBM with your traces and
monitors. :)

What was the PTF/APAR?

Charles

On Tue, Nov 21, 2023 at 11:49 PM Martijn van Breden <
m.vanbreden@xxxxxxxxxxxxxxxxxxxxxxxxxx<mailto:m.vanbreden@xxxxxxxxxxxxxxxxxxxxxxxxxx>> wrote:

> Hi all,
>
> Just an update on this issue to keep you informed.
>
> As some of you already mentioned it's highly unlikely that the query
> engine returns different results when exactly the same query is invoked
> from different environments. Like in my case, from within SQLRPGLE or from
> ACS-RSS.
>
> Highly unlikely, but not impossible. I ran tests for IBM support and
> uploaded a bunch of traces, monitors and so on. The query implementation on
> the background was the same, but the results weren't. We were adviced to
> install a PTF (whilst running on 7.4  TR7) which solved the problem.
>
> So what  I've learned from this. I've you're convinced that you see what
> you see, but still can't believe that you see what you see, fly in an extra
> pair of eyes...
>
>
> > ________________________________
> > Van: RPG400-L <rpg400-l-bounces@xxxxxxxxxxxxxxxxxx<mailto:rpg400-l-bounces@xxxxxxxxxxxxxxxxxx>> namens Charles Wilt
> <charles.wilt@xxxxxxxxx<mailto:charles.wilt@xxxxxxxxx>>
> > Now having said that, I'll say the above is true 99.99999<http://99.99999>% of the time.
> >
> > The other 0.00001%, a bug in the OS causes the difference in run-time
> > environments to surface different results.
> >
> > Note this isn't RPG vs ACS, it's just that the defaults settings *FIRSTIO
> > vs. *ALLIO, memory available to the QZDASOINIT job vs. interactive...ect.
> > If you configure the environments the same, you'd get the same
> right/wrong
> > answer from each tool.
> >
> > If you run a DB monitor and capture each invocation, you might be able to
> > tell where the statement is different.
> >
> > Or you've got something to hand to IBM and ask them what is going on.
> >
> > HTH,
> > Charles