I would strongly suggest the:

optimizer_features_enable = 8.1.6

because it also cures another problem with 8.1.7 wherein a complex query 
with multiple OR and AND conditions will return the wrong number of rows if 
you are using CBO (and/or any statistics are on the resultant table).

larry

At 07:55 PM 4/5/2001 -0800, you wrote:
>Hi,
>
>There is a note on Metalink (137430.1 reproduced below) describing a generic
>817 optimiser problem that will result in incorrect results when using
>certain query types.
>
>We are currently running 81511 on NT but testing an upgrade to 81711 on NT
>at the moment.
>
>The note suggests 2 workarounds, and my question to others is
>
>Should we set "optimizer_features_enable = 8.1.6"
>-or-
>should we set "_eliminate_common_subexpr = false"
>
>Which method does the list suggest, and what are the implications of each?
>
>Thanks,
>Bruce Reardon
>mailto:[EMAIL PROTECTED]
>
>
>
>Doc ID:  Note:137430.1
>results
>Type:  ALERT
>Status:  PUBLISHED
>  Content Type:  TEXT/PLAIN
>Creation Date:  23-MAR-2001
>Last Revision Date:  28-MAR-2001
>
>
>   Common Subquery Elimination in 8.1.7 causes incorrect results
>   ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>   Versions Affected
>   ~~~~~~~~~~~~~~~~~
>   Oracle Server releases 8.1.7.0 and 8.1.7.1
>
>   Fixed in 8.1.7.2 and 9.0.0.0
>
>   Platforms Affected
>   ~~~~~~~~~~~~~~~~~~
>     GENERIC
>
>   Description
>   ~~~~~~~~~~~
>   The bug is that rows may be returned when none are expected with the query
>below.
>   Other queries may return more or less rows than expected depending on the
>predicates used.
>
>
>     select * from tab1
>      where (tab1.col1 = 1 and exists (select * from tab2))
>         or (tab1.col1 = 2 and exists (select * from tab3))"
>
>
>There are several variations on this.
>Eg:
>select /*+ ALL_ROWS */ * from tab1
>       where (tab1.col1 = 1 and col1 in (select col1 from tab2))
>          or (tab1.col1 = 2 and col1 in (select col1 from tab3))
>;
>
>
>
>
>   Note:  The use of the two parts of the WHERE clause separated by an OR
>operator,
>   each clause containing a non-correlated subquery.
>
>
>   The problem is that common subquery elimination treats un-correlated
>subqueries
>   as identical even when they are not.
>
>
>
>   Likelihood of Occurrence
>   ~~~~~~~~~~~~~~~~~~~~~~~~
>
>   Any time the above construct is used.
>   This construct is common in queries used by some third party applications
>including
>   People Soft.
>
>   This problem only occurs on database versions 8.1.7.0 and 8.1.7.1.
>   Earlier versions are not affected.
>
>
>   Possible Symptoms
>   ~~~~~~~~~~~~~~~~~
>   No error is raised.
>   Unless the user is aware of incorrect results the problem will be
>undetected.
>
>
>
>
>   Workaround
>   ~~~~~~~~~~
>   As the problem is caused by a new optimizer feature introduced in 8.1.7,
>   the workaround is to set a parameter in the init<SID>.ora to disable the
>feature.
>
>
>Either set:
>
>
>   "optimizer_features_enable = 8.1.6"
>   -or-
>   "_eliminate_common_subexpr = false"
>
>
>   Patches
>   ~~~~~~~
>   This is expected to be fixed in the 8.1.7.2 Patch Set.
>   The problem is fixed in Oracle9i.
>   No single one off patches are currently available.
>
>
>
>
>   References
>   ~~~~~~~~~~
>      QUERY RETURNING ROWS IN 8.1.7 WHEN NO ROWS EXPECTED    [BUG:1578644]
>--
>Please see the official ORACLE-L FAQ: http://www.orafaq.com
>--
>Author: Reardon, Bruce (CALBBAY)
>   INET: [EMAIL PROTECTED]
>
>Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
>San Diego, California        -- Public Internet access / Mailing Lists
>--------------------------------------------------------------------
>To REMOVE yourself from this mailing list, send an E-Mail message
>to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
>the message BODY, include a line containing: UNSUB ORACLE-L
>(or the name of mailing list you want to be removed from).  You may
>also send the HELP command for other information (like subscribing).

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: Larry Strickland
  INET: [EMAIL PROTECTED]

Fat City Network Services    -- (858) 538-5051  FAX: (858) 538-5051
San Diego, California        -- Public Internet access / Mailing Lists
--------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).

Reply via email to