perfomance problem - Mailing list pgsql-general

Hi everybody,

I'm having some trouble optimizing the performance of a query on my web log system. The table is 1,6 millions records ( secu.logs_et table).

Here is the query :
------------------------------------------------------------------------------------
SELECT profiles_et.username, profiles_et.name, profiles_et.firstname, profiles_et.email, company_et.name AS company, count(logs_et.dt) AS cnt, pro_invoice_addr_at.id


FROM
(
     (
                 (
                           secu.company_et  JOIN secu.pro_invoice_addr_at ON (pro_invoice_addr_at.company = company_et.id)
                 )
              
               JOIN secu.profiles_et ON (pro_invoice_addr_at.profile = profiles_et.id)
     )
    JOIN  secu.logs_et ON (logs_et.invaddr = pro_invoice_addr_at.id)
)


 GROUP BY profiles_et.username, profiles_et.name, profiles_et.firstname, profiles_et.email, company_et.name, pro_invoice_addr_at.id

ORDER BY count(logs_et.dt) DESC;
------------------------------------------------------------------------------------



The secu_company, secu_prov_invoice_addr_at and secu_profiles contains only informations about the users.

The query is executing in 8,6 minutes.... Indexes are defined on all the field used in the different joins.

Here is the query plan :

------------------------------------------------------------------------------------
Sort  (cost=895649.54..896073.23 rows=169474 width=145)
  Sort Key: count(logs_et.dt)
  ->  Aggregate  (cost=831240.24..865135.10 rows=169474 width=145)
        ->  Group  (cost=831240.24..860898.24 rows=1694743 width=145)
              ->  Sort  (cost=831240.24..835477.10 rows=1694743 width=145)
                    Sort Key: profiles_et.username, profiles_et.name, profiles_et.firstname, profiles_et.email, company_et.name, pro_invoice_addr_at.id
                    ->  Merge Join  (cost=274406.73..304066.75 rows=1694743 width=145)
                          Merge Cond: ("outer".id = "inner".invaddr)
                          ->  Sort  (cost=143.81..144.92 rows=446 width=126)
                                Sort Key: pro_invoice_addr_at.id
                                ->  Merge Join  (cost=90.27..124.18 rows=446 width=126)
                                      Merge Cond: ("outer".id = "inner".profile)
                                      ->  Index Scan using profiles_pk on profiles_et  (cost=0.00..24.98 rows=449 width=66)
                                      ->  Sort  (cost=90.27..91.39 rows=446 width=60)
                                            Sort Key: pro_invoice_addr_at.profile
                                            ->  Merge Join  (cost=37.82..70.65 rows=446 width=60)
                                                  Merge Cond: ("outer".company = "inner".id)
                                                  ->  Index Scan using invaddr_at_company_idx on pro_invoice_addr_at  (cost=0.00..24.68 rows=446 width=33)
                                                  ->  Sort  (cost=37.82..38.55 rows=291 width=27)
                                                        Sort Key: company_et.id
                                                        ->  Seq Scan on company_et  (cost=0.00..25.91 rows=291 width=27)
                          ->  Sort  (cost=274262.92..278499.78 rows=1694743 width=19)
                                Sort Key: logs_et.invaddr
                                ->  Seq Scan on logs_et  (cost=0.00..55404.43 rows=1694743 width=19)
------------------------------------------------------------------------------------

The computer is a Pentium III 850 Mhz with 256 mb RAM (Sort memory for postres is 16 mb, and 64 mb of shared memory).


Do someone have any idea on how to speed up the query ? I can give any more details about the system if needed.

Thanks a lot in advance

Didrik Pinte

,"**** DISCLAIMER ****
This e-mail and any attachments thereto may contain information
which is confidential and/or protected by intellectual property
rights and are intended for the sole use of the recipient(s) named above.
Any use of the information contained herein (including, but not limited to,
total or partial reproduction, communication or distribution in any form)
by persons other than the designated recipient(s) is prohibited.
If you have received this e-mail in error, please notify the sender
and delete the material from any computer.
Thank you for your cooperation.