:[EMAIL PROTECTED] On Behalf Of
Sai Selvaganesan
Sent: 28 May 2003 06:15
To: Multiple recipients of list ORACLE-L
Subject: Re: 10053 trace
hi
thanks for the info.i could make it work for some 9i
sqls but a particular sql is not generating trace.
the 10053 trace file has only the query
:
hi
thanks for the info.i could make it work for some 9i
sqls but a particular sql is not generating trace.
the 10053 trace file has only the query para but no
other details. the query selects from a simple view
built on a table. will this make a difference?
here is the query
select id from
hi gurus
i am trying to generate 10053 trace in 9i and have met
with no success.
i used both 'alter session set events' and
'dbms_system.set_ev'.
i used the above two to generate 10053 trace for the
same query in 8i w/o problems.
has anything changed in 9i?
can you please help me
It has worked for me in 9i as well.
What is it that you are not getting?
Please review Wolfgang's 10053 Events paper at
http://www.centrexcc.com. He may have any 9i specific
information.
- Kirti
--- Sai Selvaganesan [EMAIL PROTECTED] wrote:
hi gurus
i am trying to generate 10053 trace
At 04:15 PM 5/27/2003 -0800, you wrote:
hi gurus
i am trying to generate 10053 trace in 9i and have met
with no success.
i used both 'alter session set events' and
'dbms_system.set_ev'.
i used the above two to generate 10053 trace for the
same query in 8i w/o problems.
has anything changed in 9i
hi
thanks for the info.i could make it work for some 9i
sqls but a particular sql is not generating trace.
the 10053 trace file has only the query para but no
other details. the query selects from a simple view
built on a table. will this make a difference?
here is the query
select id from
, thus ensuring that the CBO will need to
parse. If the CBO doesn't need to parse, then event 10053 doesn't generate
trace output...
Hope this helps...
-Tim
on 5/27/03 5:15 PM, Sai Selvaganesan at [EMAIL PROTECTED] wrote:
hi gurus
i am trying to generate 10053 trace in 9i and have met
I'm looking at the results of a 10053 trace from 9i.
I'm not finding any real documentation for the following parameters in the
base table access cost section:
tb_sel
rsc_cpu
rsc_io
ix_sel
I believe that these all combine somehow to define a cost for a specific
index
access but I'm not sure how
Shao:
I think your program has become a victim of the new costing model
for bitmap indexes which was introduced in 8.0.6. Basically the
new costing model assumes the 80% of the rows are costed in the
old model and 20% of the rows are costed as they are spread across
ALL blocks.
Let us assume
I have a query in which the predictate is like this:
WHERE ln.metro_id = i_metro_id AND
l.metro_id = :b1 ANd
eld.metro_id(+) = :b1 AND
ln.status_desc = 'ACTIVE' AND
ln.daset_id = ld.daset_id (+) AND
10 matches
Mail list logo