I did not follow the discussion, but I would say you need to have a combined 
index on a.timestamp, a.a_id and a.c_id. Sorry if you already tried this.
Stefan

Am Friday 30 July 2004 21:34 schrieb Eamon Daly:
> So, to confirm, short of indexing a.timestamp (which I've
> done) there's no way to speed up the original query?
>
> SELECT *
> FROM a
> LEFT JOIN b ON a.a_id = b.a_id
> JOIN c ON a.c_id = c.c_id
> JOIN d ON c.d_id = d.d_id
> JOIN e ON c.e_id = e.e_id
> WHERE a.timestamp BETWEEN 20040101000000 AND 20040101235959
> GROUP BY c.d_id, c.e_id
>
> This report takes over an hour when looking at just one
> month's worth of data. It's brutal. Any and all suggestions
> would be appreciated.
>
> ____________________________________________________________
> Eamon Daly
>
>
>
> ----- Original Message -----
> From: "gerald_clark" <[EMAIL PROTECTED]>
> To: "Eamon Daly" <[EMAIL PROTECTED]>
> Cc: <[EMAIL PROTECTED]>
> Sent: Thursday, July 29, 2004 10:54 AM
> Subject: Re: JOIN/WHERE and index confusion
>
> > No.  a.a_id is used in the join to b, not for selecting records in a.
> >
> > Eamon Daly wrote:
> > >I would assume it would use reporting_id_t, since the WHERE
> > >clause has both a.a_id and a.timestamp in it.
> > >
> > >____________________________________________________________
> > >Eamon Daly
> > >
> > >----- Original Message -----
> > >From: "gerald_clark" <[EMAIL PROTECTED]>
> > >To: "Eamon Daly" <[EMAIL PROTECTED]>
> > >Cc: <[EMAIL PROTECTED]>
> > >Sent: Thursday, July 29, 2004 8:04 AM
> > >Subject: Re: JOIN/WHERE and index confusion
> > >
> > >>Why should it use any other?
> > >>Timestamp is the only field in the where clause, and you are selecting
> > >>all fields.
> > >>The second index would require more index reads, and the third can't be
> > >>used
> > >>to satisfy the where clause.
> > >>
> > >>Eamon Daly wrote:
> > >>>Okay, now I'm even /more/ confused. I whittled everything
> > >>>down like so:
> > >>>
> > >>>CREATE INDEX reporting_t ON a (timestamp);
> > >>>CREATE INDEX reporting_t_id ON a (timestamp, a_id);
> > >>>CREATE INDEX reporting_id_t ON a (a_id, timestamp);
> > >>>
> > >>>EXPLAIN
> > >>>SELECT *
> > >>>FROM a, b
> > >>>WHERE
> > >>>a.a_id = b.a_id AND
> > >>>a.timestamp BETWEEN 20040101000000 AND 20040101235959
> > >>>
> > >>>and it /still/ only uses reporting_t! What the heck am I
> > >>>missing?
> > >>>
> > >>>____________________________________________________________
> > >>>Eamon Daly
> > >>>
> > >>>----- Original Message -----
> > >>>From: "Eamon Daly" <[EMAIL PROTECTED]>
> > >>>To: <[EMAIL PROTECTED]>
> > >>>Sent: Wednesday, July 28, 2004 10:58 AM
> > >>>Subject: JOIN/WHERE and index confusion
> > >>>
> > >>>>Hi, all. I think I'm misunderstanding something basic about
> > >>>>indexes. I have a SELECT like so:
> > >>>>
> > >>>>SELECT *
> > >>>>FROM a
> > >>>>LEFT JOIN b ON a.a_id = b.a_id
> > >>>>JOIN c ON a.c_id = c.c_id
> > >>>>JOIN d ON c.d_id = d.d_id
> > >>>>JOIN e ON c.e_id = e.e_id
> > >>>>WHERE a.timestamp BETWEEN 20040101000000 AND 20040101235959
> > >>>>GROUP BY c.d_id, c.e_id
> > >>>>
> > >>>>All of the id fields are primary indexes. I've already
> > >>>>created an index on a.timestamp, and that works all right.
> > >>>>I tried creating an index on a for the SELECT:
> > >>>>
> > >>>>KEY `reporting` (`a_id`,`c_id`,`timestamp`)
> > >>>>
> > >>>>and an index on c for the GROUP BY:
> > >>>>
> > >>>>KEY `reporting` (`c_id`,`d_id`,`e_id`)
> > >>>>
> > >>>>But EXPLAIN shows that MySQL isn't even considering the key
> > >>>>on a, and chooses the primary key on c over my index.
> > >>>>Clearly I'm confused about how indexes are used in a
> > >>>>JOIN/WHERE situation: can anyone enlighten me?
> > >>>>
> > >>>>____________________________________________________________
> > >>>>Eamon Daly

-- 
Stefan Kuhn M. A.
Cologne University BioInformatics Center (http://www.cubic.uni-koeln.de)
Zülpicher Str. 47, 50674 Cologne
Tel: +49(0)221-470-7428   Fax: +49 (0) 221-470-7786
My public PGP key is available at http://pgp.mit.edu


-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:    http://lists.mysql.com/[EMAIL PROTECTED]

Reply via email to