Re: Possible SQL Activation Group Issues

"MIDRANGE-L" <midrange-l-bounces@xxxxxxxxxxxx> wrote on 10/14/2016
02:05:07 PM:

It seems strange that they would tell you to make sure you ran in the

same

activation group and to make the close cursor behaviour *ENDACTGRP. That
seems like the opposite of making sure that each run of the RPG is data
path unique. For example, wouldn't *ENDMOD close any open SQL cursors

more

effectively at the end of each RPG run? Even further assurance would be

to

run in a *NEW activation group.

My batch application is using an MVC architecture.

CLLE --> non-SQL RPGLE "Controller" (with PRTF "View") --> SQLRPG "Service
Layer" --> SQLRPG DAO "Model"

Thus, we can't use *ENDMOD because we would lose the cursor which
only exists in the Service Layer.

Also, what was the point of them telling you to use dynamic SQL if you
aren't already specifying the library in the prepared statement. Okay,
perhaps this one. I know that data paths are optimized on the RPG

program

object itself, so maybe they were trying to avoid some reuse there, but

I'm

kind of ignorant on how that works.

We did not want to be qualifying our table and view names in
programs. We wanted the library list to control what is accessed. They
said that under the conditions specified that dynamic SQL will tend to use
the library list where static SQL will tend to use only the current schema
-- which defaults to the name of the current user. We also did not want
to be setting schema names in programs. Again, we wanted the library list
to control what is being accessed.

Sincerely,

Dave Clark