Ciao Dario,
On Tue, 21 Dec 2010 09:14:36 +0000, Dario Beraldi
<dario.beraldi@ed.ac.uk> wrote:
> the query run better the planner is not able (allowed?) to create
> such
> index, use it, and drop it once the query is done. Why is it so?
Because it is not its responsibility. This is the simplest and most
rational answer.
I would reply with some questions that hopefully will give you the
answer. How can you define a 'good' query from a 'bad' query? Consider
the case when an user launches a wrong query and the planner you propose
starts creating an index. I believe that would be a mess.
Then ... how could you project this scenario in a concurrent context
where multiple users launch queries that 'need' an index?
I suggest that you look at the documentation for more information.
Otherwise, I strongly suggest that you read the chapter on the planner
from Greg's book on High Performance (which you can find from here:
http://www.postgresql.org/docs/books/)
Merry Christmas to you too!
Cheers,
Gabriele
--
Gabriele Bartolini - 2ndQuadrant Italia
PostgreSQL Training, Services and Support
Gabriele.Bartolini@2ndQuadrant.it - www.2ndQuadrant.it
