Re: Microsoft Copilot and IBM i DB2 SQL

[Richard Schoen <richard@xxxxxxxxxxxxxxxxx>]

Yep, describe it to AI like it's a 5yr old 😊

Definitely mention IBMi and DB2 and other specific info that's detailed. 

I also give it some context description as well.

Prompt engineering is interesting. The more specific the better. 

I did an example yesterday where I had over 100 fields in a data structure and it saved me a couple hours by just turning that into a JSON_TABLE statement.

Regards,
Richard Schoen
Web: http://www.richardschoen.net
Email: richard@xxxxxxxxxxxxxxxxx

------------------------------

message: 4
date: Wed, 18 Mar 2026 12:11:43 -0700
from: Arne Quanbeck <arne@xxxxxxxxxxxx>
subject: Re: Microsoft Copilot and IBM i DB2 SQL

Generalized LLMs like Grok and ChatGPT need to know what system they are producing results for. Stephen's advice is excellent. You can add additional preferences to get a better solution. 

The other thing to note is that there are generally multiple versions of each model. For Grok the current generally available version as of this writing is 4.2. Grok has Fast, Expert and Heavy modes. You need a subscription for heavy, and it's generally overkill for things like generating SQL.


On Wed, 2026-03-18 at 15:01 -0400, Stephen Richter wrote:
> Maybe you need a subscription??? I pay $12 a? month through x.com .??? 
> Post your question, I will run it through grok.?? In the prompt I 
> always say DB2 for IBM i.? The SQL answers are outstanding from Grok.
>
> On Wed, Mar 18, 2026 at 2:41?PM Art Tostaine, Jr.
> <atostaine@xxxxxxxxx>
> wrote:
>
> > Just today Grok didn't have any idea what i was talking about, 
> > ChatGPT gave me a perfect SQL statement with the exact same prompt.
> >
> > Art
> > --
> > This is the Midrange Systems Technical Discussion (MIDRANGE-L) 
> > mailing list To post a message email: MIDRANGE-L@xxxxxxxxxxxxxxxxxx 
> > To subscribe, unsubscribe, or change list options,
> > visit: https://lists.midrange.com/mailman/listinfo/midrange-l
> > or email: MIDRANGE-L-request@xxxxxxxxxxxxxxxxxx
> > Before posting, please take a moment to review the archives at 
> > https://archive.midrange.com/midrange-l.
> >
> > Please contact support@xxxxxxxxxxxxxxxxxxxx?for any subscription 
> > related questions.