Re: Oracle versus MS Sql Server

"Howard J. Rogers" <howardjr_at_www.com> schrieb im Newsbeitrag news:3bdb1977$0$8347$afc38c87_at_news.optusnet.com.au...
> I rest my case m'lud.
>
> Verdict: he hasn't a clue.
>
> HJR
Come on. This was a joke.

Of course I also could have cited some renowned expert as for example "HJR, 2001-03-02, comp.databases.oracle.server":

>>>

There are various formulae out there designed to work out PCTFREE... all of them are somewhat dodgy, so I ignore them and look at my table definitions. If there are few NOT NULLS, PCTFREE should be higher. If there are lots of VARCHAR2s and NUMBER fields, PCTFREE should be higher. If a significant proportion of the record is prone to updates, PCTFREE should be higher. Higher than what? The default of 10%. A table, subject to updates, with few not nulls, and comprised of a lot of varchar2s and number fields should be a candidate for 25% or so. Particularly if there are few full table scans on such a table, I'd err on the side of higher PCTFREEs and who cares at the waste of space and low data density in the blocks: if access is largely via indexes, then it is more important to ensure that the indexes point to the right places first time. Tables subject to lots of full tablescans need to be handled with more care -where the indexes points to is less critical (since they're not being used!), and you want to retrieve as much data in as few block reads as possible. <<<

However, there is not the "flies thing" in it.

This thread is really out of control. I will read any answers from you, but for saving your time and my time as well as storage capacity on the server, I won't answer - except if you ask me to do so.

It's been interesting. I did learn some things. Maybe not too many new friends. But some parts were good for understanding the "other side". Have a nice day.

Michael G. Schneider Received on Sat Oct 27 2001 - 16:14:41 CDT