Re: Views of views, complexity and speed.

El martes 05 de febrero, Jan Wieck escribió:
> bombadil(at)wanadoo(dot)es wrote:
> > In order to avoid complexity, some views looks in other views and
> > join them for getting data.
> >
> > I see queries against that views result slower than queries against
> > plane tables or simple views by an order of magnitude (when not two).
> >
> > My question is: if I would make complex views looking in plain tables
> > instead of other views, could I gain speed with the cost of more
> > difficult maintainability and readability?
> >
> > Sorry for lazy data and arguments. If any of you think that detailed
> > tables and views may help, i can send them without problem.
>
> Asking for qualified opinions and comments "only" and then
> beeing lazy on data and arguments, tztztz ... man!

Emmmmm, sorry 0:)

> The question I have is what do you really compare? You said
> "looking in plain tables instead of other views". Does that
> mean your query is faster when you build one big view against
> all the base tables instead of cascaded views, or what? What
> is the performance difference if you instead of using the
> cascaded views query all the base tables in a big join
> directly?

Your comment resumes very well my essential question. I only want to
know if there is a reason for thinkink that a cascade of views can be
slower than a complex view that includes all tables. Is the planner
well tuned for working with this complex cases (cascade of views)?

Actually I am making experiments with all this stuff. Better idea
than sending a lazy question in the list.

Sorry again, and thanks for interest.

Greets.

David