Re: Strictly Enforcing Query Plan Stability

The schema is OUTLN - the objects are not owned by SYS - so you may as well try to create the trigger.

It is possible that the certain but not all errors are trapped by the 'recursive sql' of create ouline - after all Oracle, outline creation has to fail sometimes without stopping the query or Oracle would be in trouble. Look in oraus.msg to see if there are any errors that look like - 'cannot continue, not allowed to create/access outline' and create, then raise, an exception for them in your trigger.

--

Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases
Publishers:  Addison Wesley Longman
Book bound date now 1st Dec 2000

Greg Stark wrote in message
<87k8aqmdtg.fsf_at_HSE-MTL-ppp62193.qc.sympatico.ca>...


>"Howard J. Rogers" <howardjr_at_www.com> writes:

>

>> I was under the impression that you can't place triggers on objects owned

 by


>> SYS.

>

>Well I just tried making a constraint on OUTLN.OL$ for CREATOR <> 'STARK'.

 It


>successfully blocked outlines from being created by the STARK user despite

>create_stored_outlines being set. However the queries themselves still ran.

>The error that the recursive SQL presumably caused didn't cause any error

 in


>the original SQL.

>

>I haven't tried triggers but presumably the same effect would occur if you

>could create the trigger at all.

>

>> I may be wrong.

>>

>> But if not, it was a cunning plan you had there.  Unfortunately, not as

>> cunning as a professor of cunning from Oxford University.

>

>So cunning you could brush your teeth with it.

>

>--

>greg