AW: SQL Cursor Position Problem in V5R3?

["Hauser, Birgitta" <Birgitta.Hauser@xxxxxxxxxxx>]

Hi,

> Prior to V5R3 the program worked fine.  Now, when they run
> it the first time, it works fine, but the next time they
> run it it produces no results, as if it started reading at
> the EOF.

Did you define a serial or a scroll cursor? 
If you use a serial cursor, you have to close it at the end, otherwise it
cannot be repositioned, that means a new OPEN will have no effect.
To make sure, that the cursor is really closed, I normally close it
immediatedly before the OPEN statement.

A new activation group forces hard closes of all cursors at the end of the
program. An other way to force hard closes at the end of a module is to set
the option CLOSQLCSR in the compile command to *ENDMOD. From the view of the
performance, both methods are not ideal, because an ODP (open data path)
cannot be reused in subsequent executions. 

Birgitta

Mit freundlichen Grüßen

i.A. Birgitta Hauser

LUNZER + PARTNER GMBH
Unternehmensberatung
Carl-Zeiss-Straße 1
63755 Alzenau

Tel:         + 49 6023 951-255
Fax:        + 49 6023 951-111
Internet.  www.lp-gmbh.com
              www.rpg-schulung.de

"If you think education is expensive, try ignorance"
(Derek Bok)

-----Ursprüngliche Nachricht-----
Von: midrange-l-bounces@xxxxxxxxxxxx
[mailto:midrange-l-bounces@xxxxxxxxxxxx] Im Auftrag von Barbara Morris
Gesendet: Samstag, 28. Januar 2006 01:47
An: midrange-l@xxxxxxxxxxxx
Betreff: Re: SQL Cursor Position Problem in V5R3?

Brad Stone wrote:
> 
> I am working with a customer who is having a problem with
> what seems like a cursor position problem using embedded
> SQL.
> 
> Prior to V5R3 the program worked fine.  Now, when they run
> it the first time, it works fine, but the next time they
> run it it produces no results, as if it started reading at
> the EOF.
> 

It might be an activation group difference, if the program is created
with CRTPGM and the ACTGRP parameter is not specified.

Prior to V5R3, a simple CRTPGM always created a program in *NEW
activation group.  In V5R3 it was changed so that the default value used
for the ACTGRP parameter depends on the type of program-entry module. 
For RPGLE modules, the default is QILE, not *NEW.