Darren King wrote:
>
> Bruce,
>
> The "SELECT on two tables where zero or one table in WHERE or target
> clause returns no rows" appears to have been fixed in 6.1. I'll put
> the script that I tested it with at the end of this message. Works
> for me.
>
> Should I put a test like this into the regression tests to make sure no
> future coding breaks this?
>
> Also, could you remove the item for me to "Improve optimizer plan choice"?
May be. But there are still problems with selectivity functions.
Vadim
------------------------------