Re: Deadly sins againts database performance/scalability

From: Frank <fbortel_at_nescape.net>
Date: Fri, 28 Nov 2003 12:46:15 +0100
Message-ID: <bq7c2a$n7c$1@news3.tilbu1.nb.home.nl>

Cris Carampa wrote:

> Alkos wrote:
>

>> Most of the time, people use SELECT DISTINCT to get unique rows because
>> they don't want to "bother" with the datamodel. They are lazy or 
>> hurried so
>> they don't care about possible
>> keys, about WHERE clauses selectivity and all that kind of things so 
>> quick
>> and dirty solution to be sure to get unique rows : SELECT DISTINCT which
>> implies a sort therefore more activity to get the same result as they may
>> have gotten quicker with a little bit work on the WHERE clause.

>
>
> I do use SELECT DISTINCT when I have just to pick up distinct values
> from a table, and I don't need to count them or performing group
> functions amongs columns in the table. I assume the performance of this
> statement:
>
> SELECT DISTINCT FOOCOL FROM FOOTAB ;
>
> is better than the performance of this one:
>
> SELECT FOOCOL, COUNT(FOOCOL) FROM FOOTAB GROUP BY FOOCOL ;
>
> Or am I wrong?
>

Consider

select distinct e.first_name
from employees e, departments d;

versus

select e.first_name
from employees e, departments d
where e.department_id=d.department_id;

Oversimplified, of course, but it's the idea.

-- 
Regards, Frank van Bortel

Received on Fri Nov 28 2003 - 05:46:15 CST