Обсуждение: Explain returns extremely high cost

Поиск
Список
Период
Сортировка

Explain returns extremely high cost

От
"Andrii Vasyliev"
Дата:
Hi!

I have a database just recreated from scratch with SQL script.
(My PostgreSQL version is 8.1.3).
All of the tables are pretty empty, but that's what I've got with explain:

EXPLAIN SELECT * FROM account;
                               QUERY PLAN
------------------------------------------------------------------------
 Seq Scan on account  (cost=100000000.00..100000001.05 rows=5 width=41)

This is what I have in the table:

SELECT * FROM account;
 obj_id  |  login  |  password   | client_id | service_id | type_id | state_id
---------+---------+-------------+-----------+------------+---------+----------
 1000513 | root    | _system____ |   1000471 |    1000486 | 1000113 |  1000209
 1000548 | root    | _system____ |   1000472 |    1000536 | 1000113 |  1000209
 1000550 | root    | _system____ |   1000475 |    1000534 | 1000113 |  1000209
 1000552 | sol     | _sol        |   1000472 |    1000536 | 1000113 |  1000209
 1000554 | seregik | _seregik    |   1000475 |    1000534 | 1000113 |  1000209
(5 rows)
Time: 1.346 ms

For other tables I have more or less the same high costs.

I've done VACUUM ANALYZE. This is what I have in pg_class table:

SELECT relpages,reltuples FROM pg_class WHERE relname='account';
relpages | reltuples
----------+-----------
        1 |         5

Why the costs are so high? Is it a bug?
Thanks in advance for any suggestions!
Regards, Andrii

Re: Explain returns extremely high cost

От
Martijn van Oosterhout
Дата:
On Tue, Aug 15, 2006 at 10:14:49AM +0300, Andrii Vasyliev wrote:
> Hi!
>
> I have a database just recreated from scratch with SQL script.
> (My PostgreSQL version is 8.1.3).
> All of the tables are pretty empty, but that's what I've got with explain:
>
> EXPLAIN SELECT * FROM account;
>                               QUERY PLAN
> ------------------------------------------------------------------------
> Seq Scan on account  (cost=100000000.00..100000001.05 rows=5 width=41)

Looks like someone did an "enable_seqscan=no" somewhere, do a "show
all" and check the planner variables...

Have a nice day,
--
Martijn van Oosterhout   <kleptog@svana.org>   http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.
Вложения