0326

[Gortner, Tim]

All: Our current IDMS manual:

CA-IDMS(R)
DML Reference -- COBOL
with FORTRAN and RPG II Supplements
Release 14.0
R005/&F0CBE

..states that an 0326 error-status is possible for an OBTAIN NEXT WITHIN
set. We are trying to determine if this is possible and under what
circumstances. We're not checking for an 0326 and it may be causing us a
loop. The last DML command to execute was this OBTAIN statement:

OBTAIN NEXT PRPAVL-M WITHIN PRPRTY-PRPAVL-M.

and the ERROR-STATUS in subschema-ctrl is 0326.

FYI: Our set definition from the Subschema listing:

SET...... PRPRTY-PRPAVL-M MODE INDEX ORDER SORTED USES
SYMBOLIC INDEX #PRPRTY-PRPAVL-M

The text from the manual:
0326 - Either the record or SPF index entry cannot be found or no more
duplicates exist for the named record.

TIA,

Tim Gortner
Galileo International
Vendor Systems Development
(303) 397-5249
Tim.G...@Galileo.com

[JB Moore]

It could indicate corruption in your database (the infamous "broken chain").
You will need to dump the page with PRINT PAGE and chase the prefixes.
Jim Moore

[Brian Potts]

Tim,

I'm not sure I understand what you're asking. But if the question is "can an end of set condition occur on an index set?" The anser is yes.

Out of curiosity, if you're not checking for '0326' what are you checking for?

Brian

>>> "Gortner, Tim" <Tim.G...@DEN.GALILEO.COM> 01/29/01 04:39PM >>>

[Albus, Michael]

Hi Tim,
I think that you are refering to this from the manual:

5.40.4 Status codes
After completion of the FIND/OBTAIN WITHIN SET/AREA function, the
ERROR-STATUS field in the IDMS communications block indicates the outcome
of the operation:

Then it shows a table with possible ERROR-STATUS fields and meanings. And you saw this in the table:

0326 Either the record or SPF index entry cannot be found or

no more duplicates exist for the named record.

It is true that you could get a 0326, but only in the following case:

OBTAIN record-name WITHIN set-name USING sort-key.

True for chain sets, the old SPF indexes, and the Integrated Indexes.

Mike Albus
Computer Associates
Instructional Consultant
CA-IDMS & DB2

> ----------
> From: Gortner, Tim[SMTP:Tim.G...@DEN.GALILEO.COM]
> Reply To: IDMS Public Discussion Forum
> Sent: Monday, January 29, 2001 4:39 PM
> To: IDM...@LISTSERV.IUASSN.COM
> Subject: 0326

[Albus, Michael]

Hi Jim,

You might get a 0361, not a 0326 if you have a broken chain.

Mike Albus
Computer Associates
Instructional Consultant
CA-IDMS & DB2

> ----------
> From: JB Moore[SMTP:conl...@IX.NETCOM.COM]

> Reply To: IDMS Public Discussion Forum

> Sent: Monday, January 29, 2001 7:19 PM
> To: IDM...@LISTSERV.IUASSN.COM
> Subject: Re: 0326

>
> It could indicate corruption in your database (the infamous "broken chain").
> You will need to dump the page with PRINT PAGE and chase the prefixes.
> Jim Moore
> ----- Original Message -----
> From: Gortner, Tim <Tim.G...@DEN.GALILEO.COM>
> To: <IDM...@LISTSERV.IUASSN.COM>

[Miller, Theodore]

Tim,

We have had this happen before when running a local mode retrieval report
while the CV was up. It appeared the record was deleted after the local
mode program determined its was the next record to be read and identified
its DBKEY, but prior to actually reading it. We resolved the situation by
checking for the 0326, then re-establishing currency and issuing a new
obtain next.

If nothing could have deleted the record, I would agree with everyone else,
you have broken chains. We use IDMSDBAN to check for the structural
integrity of our data bases. You might also want to figure out how the
broken chain occurred. We have not experienced a actual broken chain
because of actual IDMS system code in over 8 years.

Hope this helps.

Theodore (Ted) B. Miller
perotsystems(tm)
972.340.5475

-----Original Message-----
From: Gortner, Tim [mailto:Tim.G...@DEN.GALILEO.COM]
Sent: Monday, January 29, 2001 4:39 PM
To: IDM...@LISTSERV.IUASSN.COM

[Jim Moore]

I stand by my original statement Mike. Tim did not specifiy if his program was in update or retrieval, CV or local. If it was running in local mode in retrieval against a DB area that was being updated by the CV, maybe what he is encountering are "spurious" or false broken chains. What's the SQL term for this? Transient reads?

Jim Moore

IDMS Public Discussion Forum <IDM...@LISTSERV.IUASSN.COM> wrote:
> Hi Jim,

You might get a 0361, not a 0326 if you have a broken chain.

Mike Albus
Computer Associates
Instructional Consultant
CA-IDMS & DB2

> ----------
> From: JB Moore[SMTP:conl...@IX.NETCOM.COM]
> Reply To: IDMS Public Discussion Forum
> Sent: Monday, January 29, 2001 7:19 PM
> To: IDM...@LISTSERV.IUASSN.COM
> Subject: Re: 0326
>
> It could indicate corruption in your database (the infamous "broken chain").
> You will need to dump the page with PRINT PAGE and chase the prefixes.
> Jim Moore
> ----- Original Message -----
> From: Gortner, Tim