scott wrote:
> I guess if the intermediate result somewhere 
> consists of almost the whole table it causes some slowdown...

That it will do, yes!  I suspect "unique" and "group by" are doing things 
differently, with "group by" accumulating results and "unique" 
post-processing them.

You definitely want a realistic data set before you worry about speed.

-- 
   Darren New, San Diego CA, USA (PST)
   My fortune cookie said, "You will soon be
   unable to read this, even at arm's length."

Post a reply to this message