RE: Failed SETLL leaving program at undefined position in SQL VIEW, butnot in DDS LF

I'm probably wrong, but I thought that *Start positioned the pointer to the first record in the file regardless of whether or not it was indexed. I.e., to the first available (undeleted) record in file. I thought I had used it on unkeyed files before, but I could not find an example quickly so maybe Jethro Tull is making me hallucinate.

Jerry C. Adams
IBM System i Programmer/Analyst
--
B&W Wholesale
office: 615-995-7024
email: jerry@xxxxxxxxxxxxxxx


-----Original Message-----
From: rpg400-l-bounces@xxxxxxxxxxxx [mailto:rpg400-l-bounces@xxxxxxxxxxxx] On Behalf Of Carel Teijgeler
Sent: Thursday, March 18, 2010 3:23 PM
To: rpg400-l@xxxxxxxxxxxx
Subject: RE: Failed SETLL leaving program at undefined position in SQL VIEW, butnot in DDS LF

Tu use *START the file should be keyed and this is a programme described
file. Won't do, unfortunately.

With regards,
Carel Teijgeler

*********** REPLY SEPARATOR ***********

On 18-3-2010 at 13:55 DeLong, Eric wrote:

James, I doubt this applies to MI, but in RPGLE I'd use the special value

*START. In general, this is why I avoid file I/O by RRN.

-Eric

-----Original Message-----

This is really about an MI program (see the thread on the MI list), but it

also happens in an RPG test program that approximates the

point-of-failure in the MI, so maybe somebody who's here, but not on the

MI list might have an insight.

It seems that with this code (intentionally opening the file in

program-described mode, as that's essentially what happens in the

original MI program):

FHPADLPUA IF F 1243 DISK
D RECORD DS 1243
C 1 SETLL HPADLPUA

121314

C READ HPADLPUA RECORD

if I run it on a DDS LF, and RRN 1 is omitted by a SELECT clause in the

DDS, the SETLL comes back with an error, but the READ still finds the

first record that satisfies the SELECT.

But if I run an otherwise identical program on an equivalent SQL VIEW of

the same data, the READ fails, evidently from the file being left at an

undefined position.

SETGT (and this surprised me) also returned with an error, and left the

SQL VIEW at an undefined position.

Anybody know of a way, short of iteratively SETLLing until a record is

found, to deal with this?