Ian
Not sure its related, but I had a similar problem recently on 8.1.7.0 - a
query returning the wrong number of rows - but in this case it was throwing
an ora-7445 after a few reruns and the table had both function-based and IMT
indexes. OWS came up with 'Stack trace matches bug 1561106 which
Any chance its doing a INDEX DESC or INDEX MAX/MIN in
the plan? They're are two culprits I've seen in the
past that cause 'peculiar' result set to come back
hth
connor
--- MacGregor, Ian A. [EMAIL PROTECTED]
wrote: Accounts payable. But this is not a
Peoplesoft
problem, but an Oracle one.
Title: RE: Wrong Results Bug in Oracle 8.1.7.1
We had a similar issue, although it always resulted in an ORA-600. It was with descending indexes as defined in the Peopletools repository. We had to set _IGNORE_DESC_IN_INDEX=TRUE (and _DB_FILE_NONCONTIG_MBLOCK_READ_COUNT=1 to comply
We had this same problem last year. The answer was to
upgrade to 8.1.7.2 which came with its own set of bugs
which required we go to 8.1.7.3.
We had run for months in production before we ran into
the bug. Once we hit it, there was no way to avoid
it. The only option was to upgrade.
I don't
These types of problems can sometimes be resolved with an index rebuild,
btw.
hth,
Paul
---
www.pythian.com -- [EMAIL PROTECTED] -- 877-PYTHIAN
Smarter than adding another team member, Pythian has new services for
supplementing DBAs: get our help with monitoring, 24x7 on-call, daily
I recall a similar situation with Oracle 7.3 and parallel index creations
where the row source and output would get reversed. Index scans would return
now rows (The developer called with 'I just created and index and now the
data is all gone'). If a FTS works, but not the index, that tells me
Title: RE: Wrong Results Bug in Oracle 8.1.7.1
That's
it. The init.ora parameter, IGNORE_IN_INDEX had been
set to TRUE, but the indexes had not been rebuilt
Ian
MacGregor
Stanford Linear Accelerator Center
ian@SLAC.Stanford.edu
-Original Message-From: Aponte, Tony
[mailto
Ian, is that peopleslop?
joe
MacGregor, Ian A. wrote:
SQL SELECT VOUCHER_ID FROM PS_VCHR_MM_VW WHERE ROWNUM 10;
VOUCHER_
3394
3395
3396
3397
3398
3399
3400
3401
3402
set
Yep sure is.
Ian
-Original Message-
Sent: Sunday, June 09, 2002 5:43 AM
To: Multiple recipients of list ORACLE-L
Ian, is that peopleslop?
joe
MacGregor, Ian A. wrote:
SQL SELECT VOUCHER_ID FROM PS_VCHR_MM_VW WHERE ROWNUM 10;
VOUCHER_
3394
3395
3396
what modules, if you dont mind me asking, i'm at a site where we're
going to implement HR, Financials And EPM soon.
thanks, joe
MacGregor, Ian A. wrote:
Yep sure is.
Ian
-Original Message-
Sent: Sunday, June 09, 2002 5:43 AM
To: Multiple recipients of list ORACLE-L
Ian, is
Accounts payable. But this is not a Peoplesoft problem, but an Oracle one. I tried
the same query against another Peoplesoft instance, and it ran fine. So there's
something more than the view involved here. It's in Oracle Support's court now.
I must be living right. It's failing in
11 matches
Mail list logo