| 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: Script Question from Cary's Book
Mohammed,
You're right. I can't believe I divided by a constant in the book, =
because
the divisor varies by Oracle version. I normally use something like =
this:
my $ora_version =3D 8; # use 7, 8, 9, or 10 my %ora_resolution =3D (
7 =3D> 0.01, # Oracle version 7 publishes centiseconds
8 =3D> 0.01, # Oracle version 8 publishes centiseconds
9 =3D> 0.000001, # Oracle version 9 publishes microseconds 10 =3D> 0.000001, # Oracle version 10 publishes microseconds);
...Then I multiply by $res instead of dividing by a constant.
You may also want to modify the precision on the %*.*f specifications, =
since
there's more information available to you to the right of the decimal =
point
in version 9+.
Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com
* Nullius in verba *
Upcoming events:
- Performance Diagnosis 101: 7/20 Cleveland, 8/10 Boston, 9/14 San =
Francisco
- SQL Optimization 101: 7/26 Washington DC, 8/16 Minneapolis, 9/20 =
Hartford
- Hotsos Symposium 2005: March 6-10 Dallas
-----Original Message-----
From: oracle-l-bounce_at_freelists.org =
[mailto:oracle-l-bounce_at_freelists.org]
On Behalf Of mkb
Sent: Tuesday, July 20, 2004 4:19 PM
To: oracle-l_at_freelists.org
Subject: Script Question from Cary's Book
Folks,
Question about the script in Cary's book (Optimizing Oracle Performance). I don't know if I can cut 'n paste the script without infringing on copyright, so forgive me if I just paste the relevant couple lines.=20
The actual script is on page 96, chapter 5: Interpreting Extended SQL Trace Data so I guess only those folks who have the book could answer, right?
I ran the script against an Oracle 9i R2 trace file running on Windows 2000. My assumption is that all timings are in micro seconds (1/1000000 sec). If that is the case, should the following lines at the bottom of the script:
printf "%8.2fs %5.1f%%, %-40s\n", $ela{$_}/100,
$ela{$_}/$r*100, $_ for sort { $ela{$b}
<=3D> $ela{$a} } keys %ela;
printf "%8s- %5s- %-40s\n", "-"x8, "-"x5, "-"x40;
printf "%8.2fs %5.1f%% %-40s\n", $r/100, 100, "Total
response time";
be written as:
printf "%8.2fs %5.1f%%, %-40s\n", $ela{$_}/1000000,
$ela{$_}/$r*100, $_ for sort { $ela{$b}
<=3D> $ela{$a} } keys %ela;
printf "%8s- %5s- %-40s\n", "-"x8, "-"x5, "-"x40;
printf "%8.2fs %5.1f%% %-40s\n", $r/1000000, 100,
"Total response time";
i.e. $ela{$_}/100 becomes $ela{$_}/1000000 and $r/100 becomes $r/1000000.
Reason is I'm getting some really way off numbers like 10s' of hours for a single row update. Don't have access to the real live database, just the trace files.
So, any ideas before I make any silly conclusions?
Thanks
mohammed
=09
=09
-- 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 Tue Jul 20 2004 - 19:57:26 CDT
 |
 |