Hi there

- Is there a way to specify/force that a certain index should stay in
memory? (Suppose I have enough RAM to hold it)

- My tables never change. The data contained is constant. Is there a way
to tell this to Postgresql so that it can perform additional
optimizations? (I am aware of the possibility of index clustering.
Anything else?)

- I often perform complex queries A1, A2, A3 which are not identical but
they have in common some identical subqueries. E.g. A1 uses subquery
B1,C,D,E1, complex query A2 uses subqueries B2,C,D,E2 and so on (C,D in
common with A1, identical...). Is there a way to make Postgresql
permanently store the results of a certain number of the last subqueries
it performs, so that A2 and A3 would be greatly sped up when performed
after A1?

Thanks.

Sascha Bohnenkamp wrote:

Thanks but it's probably not what I am looking for: I would need to make
materialized views by hand before or after A1... this is not the
scenario I am thinking about.

At the time I perform A1 I cannot foresee that I will need A2, and at
the time I perform A2 I cannot foresee that I will need A3 etc. Then the
week after I perform another long list of queries that have in common
the same subqueries among them but are not in common with A1,A2,A3... so
it's a hassle to do it by hand with materialized views.
The optimization I am talking about cannot be manual. I just want that
the database remembers the last N subqueries it performed. It is strange
that there is no such optimization in postgres, that would seem an
obvious optimization to me!?
Any other relational database software has such an option?

Thanks