On Fri, 07 Jan 2005 16:17:16 +0100, Tino Wildenhain <[EMAIL PROTECTED]> wrote: > > How do you think an index would help if you do an unconditional > count(*)?
I really don't know <grin>. I don't know the inner workings of database internals but I would guess that there would be some optimized way of counting the nodes in an index tree that would be faster than sequentially going through a table.... I suppose there is no free lunch. One row, two rows, three rows, four rows, five rows.... <snore> culley ---------------------------(end of broadcast)--------------------------- TIP 4: Don't 'kill -9' the postmaster
