On 7/1/05, Ann W. Harrison <aharrison@...> wrote:

> Rick Debay wrote:
> >
> > The server. The point was that there are solutions to either calculate
> > the results in parallel, or cache them.
>
> The optimizer gets an approximate record count for calculating inexact
> costs by counting the number of pointer pages (from RDB$PAGES)
> multiplying it by the number of data pages per pointer page (fixed for
> each page size) multiplying the result by (the size of a page / byte
> length of records). It's imprecise, but "good enough".

Yes, I was talking about something like this.
And the approximate minimum could be calculated as simply the first
value in the index, without trying to materialize it.