[rescue] Mozilla Firefox

[Charles Shannon Hendrix]

Thu, 22 Apr 2004 @ 14:32 -0400, Kevin Loch said:

> Charles Shannon Hendrix wrote:
> >For that matter, much of the time you don't really need an SQL server.
> >Flat files, especially indexed, are almost always faster than using SQL.
> 
> Faster than the MySQL C API?  For all but the most trivial datasets
> I doubt it.

There is tremendous overhead on an SQL server.  

A lot of data processing functions are simple merge operations, which
can often occur near the theoretical I/O limits of your system.  You
can't get even close to that with an SQL server.

Then there are things like btree systems, some of which can even provide
SQL queries for your ad-hoc work, but otherwise can avoid the overhead.

It depends on what your mix of traditional data processing and queries
are.

> Of course the biggest obstacle to database performance is unoptimized
> schema and queries.  

Almost everywhere I have been, I found cartesian products in the
queries, and fixing them made a night and day difference.

I get tired of seeing "DBAs" writing "select distinct", when I can't get
their job because I am "not qualified".

> The worst performance is usually obtained when a DBA insists on
> academically normalized schema.

...or when he insists that you don't normalize at all.


-- 
shannon "AT" widomaker.com -- ["Trouble rather the tiger in his lair than
the sage amongst his books For to you kingdoms and their armies are mighty
and enduring,  but to him they are but toys of the moment to be overturned
by the flicking of a finger." - anonymous     ]