Re: EXPLAIN PLAN

B C Zygmunt wrote:
>
> I am going to answer my own question on this. I determined just after sending
> the previous message that the sequence of the tables is one of the key factors
> in query speed. That is, in the query below, since I am going to join lab_meas
> and lab_smp, I want them next to each other. The same for lab_smp and sample,
> and so on. (It does appear that small tables can be stuck off anywhere.)
>

It _is_ documented. See the blurb on rule vs cost optimization in the manuals.

It's generally easier to use cost based optimization, to do this you need to generate statistics on your tables and indexes (and clusters if you have any) using the ANALYSE statement. In practice estimate statistics seems to work indistinguishably from calculate and is a hell of a lot faster.

-- 
************************************************************
* Kevin Woolley *              Visual Explorer             *
*               * 3D Landscape Visualization and Modelling *
*               * Home Page: http://www.woolleysoft.co.uk  *
*               * Email    : info_at_woolleysoft.co.uk        *
************************************************************