Re: SQL Tuning

From: Daniel Morgan <damorgan_at_exxesolutions.com>
Date: Sun, 06 Jul 2003 15:04:03 -0700
Message-ID: <3F089CD3.13E59EBC_at_exxesolutions.com>


Hans Forbrich wrote:

> Daniel Morgan wrote:
>
> > I couldn't agree less.
> >
> > The CBO's decisions are intelligent and if you learn how it works, and use
> > hints, blows away the performance of the RBO.
>
> I think there is an exception to the above statement. You are making the
> assumption that the developer is actually capable of [thinking and] leveraging
> some of the non-simplistic features of the RDBMS.
>
> In my experience, the majority of developers are not aware of things like
> bit-mapped indexex, function indexes, bitmap-join indexes, partitioning, IOTs,
> etc. Most still want to reinvent stuff like intermedia & context and auditing.
> (Support for this - look at how many 'features' like XA & constraints they are
> trying to introduce at the development language level.) Thus they still tend to
> design and code like we did in the 1980's - in which case the RBO *may* still be
> valid.
>
> However, if they leverage virtually any feature that has been introduced since
> Oracle7.1, I tend to agree with your statement.

If that is true no doubt he'll be flipping burgers in the near future. Oracle is heading in directions
where experience in Oracle 6 and 7 will not take him: Hopefully you are incorrect.

--
Daniel Morgan
http://www.outreach.washington.edu/extinfo/certprog/oad/oad_crs.asp
[Quoted] damorgan_at_x.washington.edu
(replace 'x' with a 'u' to reply)
Received on Mon Jul 07 2003 - 00:04:03 CEST

Original text of this message