Re: why index is not used in query execution plan ?

Tue, 17 Sep 2002 07:49:47 +0200, Billy Verreynne said (and I quote):

> FTS is a Good Thing (tm). :-)

With the right wind behind it (corollary).

>
> It is just another tool in workbox. Used for the right job, and it works
> exceedingly well. Combine it with PQ (the two are synonomous to me) and you
> use FTS what it is intended for - processing mass amount of data in the
> shortest possible time.

Yeah. The problem is that in many texts it's explained as some form of you beaut panacea that should be used everytime there is a need for large table joins. It's not.

>
> But I agree Nuno - it must be used under the correct circumstances. Running
> FTS on a OLTP system during peak hours, is not the best of ideas. But that
> is not to say that FTS can not solve its share of problems.

Exactly. In fact, I'd say it's got nothing to do with OLTP. ANY system that has multi-user, multi-updater characteristics will suffer from a FTS hash join during anything else other than idle. And the statement itself will NOT complete faster. Try it.

-- 
Cheers
Nuno Souto
nsouto_at_optushome.com.au.nospam