| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: Implement Parallel Processing on DB Warehouse
"Flushing is harmful"
?? In response, I am flushing the shared_pool on this test instance so =
that I can get baseline timing info. I understand that if I do it as a =
matter
of course I am losing the benefit of having parsed, reusable code and =
data already in memory.
"Hint is ignored because of incorrect syntax"
That's what I get for using a bargain table book for reference. I = thought I would be able to tell when something wasn't appropriate for a = 9i database. I tried your suggestions and it worked although I didn't = get the results I wanted. Once I parallized that index I more than = quadrupled the time it took to perform that query from about .4 seconds = to 1.7 seconds. It doesn't look like this is something I can throw at = the database but will have to implement on a case by case, query by = query basis. I was probably a little unrealistic to project '1 week' to = implement parallelism <g>. I've been working on this one little test = statement nearly that long.=20
"With how many zeroes do you spell "grateful"?"
As many as you want! $0000000000000000000000000000. With zeros you can =never have enough! I appreciate the time you took to answer my question =
and, of course, would help you move furniture if called upon. It's not = likely that you would ever need Oracle advice from me. I rely on these = lists extensively for answers but try to do as much of my homework as I =can before I use up bandwidth asking for an easily obtained answer. I've = only been on this list for about a week but I believe I recognize some = of the names from the IOUG forum on Compuserve from about five years = ago.=20
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org]On Behalf Of Mladen Gogala
Sent: Tuesday, February 10, 2004 4:51 PM
To: oracle-l_at_freelists.org
Subject: Re: Implement Parallel Processing on DB Warehouse
On 02/10/2004 04:17:39 PM, "Freeman, Donald" wrote:
> SET timing ON
>=20
> ALTER SYSTEM FLUSH SHARED_POOL;
Flushing is harmful.
>=20
>=20
select /*+ parallel(tab,4) */ count(*)
from PADOHSNAP.HRA_OBS_MAP_DETAIL tab;
> COUNT(*) FROM PADOHSNAP.HRA_OBS_MAP_DETAIL;
>=20
> ********************************************************************
>=20
>=20
>=20
>=20> SORT (AGGREGATE)
> 1 0
>=20> (Cost=3D3D248 Card=3D3D4147413)
> 2 1
> INDEX (FAST FULL SCAN) OF 'PK_HRA_OBS_MAP_DETAIL' (UNIQUE) =3D
It's not going parallel because your optimizer decided that PK is =20 cheaper. You should also issue command:
alter index PADOHSNAP.PK_HRA_OBS_MAP_DETAIL parallel 4;
if it isn't already set to parallel 4. Check DBA_INDEXES.
> I am just trying to get this little test running before I try the =3D
> informatica run again. Any advice about this is welcome Any general =20
> =3D
> advice
> about what else would be helpful in my situation I would be grateful
> to =3D
> receive also.
>=20
> Thanks,
With how many zeroes do you spell "grateful"?
-- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html ----------------------------------------------------------------- ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line. -- Archives are at http://www.freelists.org/archives/oracle-l/ FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------Received on Wed Feb 11 2004 - 08:55:25 CST
 |
 |