Giorgio Valoti <giorgio_v@mac.com> writes:
> On 07/ago/08, at 17:50, Tom Lane wrote:
>> These numbers seem pretty bogus: there is hardly any scenario in
>> which a
>> full-table indexscan should be costed as significantly cheaper than a
>> seqscan. Have you put in silly values for random_page_cost?
> No,
I looked at it more closely and realized that the cost discrepancy is
from the evaluation of the function: having to evaluate a SQL or plpgsql
function 247736 times more than explains the cost estimate differential
compared to a query that involves no function call. Some experiments
here suggest that it hardly matters whether the query uses indexscan or
seqscan because the time is dominated by the function calls anyway.
regards, tom lane