Re: fate of CLUSTER command ?

On Sun, 4 Aug 2002, Christopher Kings-Lynne wrote:

> > Yes, I have always liked CLUSTER with full text searches because you are
> > usually hitting multiple rows with a single equaltiy restriction, and
> > CLUSTER puts all the hits on the same page.
> >
> > If you look in contrib/fulltextindex, you will see mention of CLUSTER in
> > the README. It may make sense to add that to your documentation.
> >
> > Also, is there any value to contrib/fulltextindex now that we have
> > contrib/tsearch?
>
> I haven't looked at tsearch yet, but I expect it's way better than
> fulltextindex. However there's more than a few of us using fulltextindex,
> so I think it will need to stay for some while. I'm working on a new
> version of it for 7.3.
>

I'm totally agre with Chris. FTI is something another thing.
FTI is good for more or less static document collection - a cost of
insert if high for inverted indices. We've developed tsearch keeping in
mind incremental update.

FTI should be faster for short queries while tsearch is better for long one.

tsearch development focused also on real IR support - language support,
indexing of specified classes of lexemes , etc. We laready have OpenFTS
which has these features, but we want to move all functionality to
tsearch.

> I can put pointers in the README about checking out tsearch...
>
> Chris
>
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 5: Have you checked our extensive FAQ?
>
> http://www.postgresql.org/users-lounge/docs/faq.html
>

Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg(at)sai(dot)msu(dot)su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83