Re: Options for poorly performing SQL
Date: Mon, 4 Feb 2013 11:03:24 -0500
Message-ID: <CAGzKQQc9_1SAuYDqPpLAFnFooVyxP8qmj=2RmzHWQm3VkbkYsA_at_mail.gmail.com>
Sandy,
Get me a SQLT XTRACT for it (215187.1) and also one from your version of the SQL. Then I will review and offer some feedback.
Cheers -- Carlos
On Mon, Feb 4, 2013 at 10:15 AM, Sandra Becker <sbecker6925_at_gmail.com>wrote:
> Oracle - EE 11.2.0.2
> OS - SUSE11
> I have a particularly egregious SQL statement that causes problems in my
> production OLTP. It's bad enough when only one session is running it, but
> at times there may be 5 to 10 sessions running it. Then I end up on a
> conference call explaining the problem is the application and unless they
> want me to kill customer sessions, I can't address the performance issues.
>
> I have been trying for 5 years to get the development team to address it.
> I have provided a properly coded statement that retrieves the identical
> results yet performs about 95 percent faster with significantly less I/0,
> along with trace data and explain plans of both the current statement and
> the optimized statement. They refuse to even look at the statement because
> it is dynamically created in the application using javaScript. Not knowing
> javaScript, I'm not sure why it makes a difference. Perhaps other
> dynamically created statements would benefit from changes here as well.
>
> Questions: Do I have any other options to corral this statement? I
> haven't used profiles and will be reading up on them this week, but would a
> profile even be a appropriate for this situation?
>
> --
> Sandy
> Transzap, Inc.
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
-- Cheers -- Carlos Sierra http://carlos-sierra.net/ -- http://www.freelists.org/webpage/oracle-lReceived on Mon Feb 04 2013 - 17:03:24 CET