On Tue, 19 Oct 2004, John Hansen wrote:
>
> A fully integrated solution would be sweet. One that combines the
> features of the archives, the search engine, and the lists themselves.
>
> You should be aware tho, that iirc the current tally is closing in on
> 300.000 emails. Does anyone have any installations with that much data
> indexed by tsearch2, who would be able to give us some indicative
> reports on query time?
>
> Mnogosearch was scrapped due to it's speed (or lack thereof), and iirc,
> tsearch2 closely resembles the same type of mechanism for full text
> searches. Just something to think about.
in principle, www.pgsql.ru/db/mw (fts.postgresql.org in past) was designed
to be an integrated solution, but time past and we have no spare time
to continue developing (we only reinstalled fts.postgresql.org).
As for tsearch2 performance, then it's slow on full archive, but reasonably
fast searching last month (default) or last year archive.
We have tested our idea using read only archive and online news archive,
and it worked very-very fast. Main idea is to have common infrastructure
(parsers, dictionaries, ranking) for inverted index (read only part) and
tsearch2 index for changing part. Running daemon (tsearchd) accepts
text for indexing, all metadata are stored in posgresql, so searching
could be constrained using metadata. Tsearchd is just a working prototype
and much work is needed.
>
> Btw, With regards to the forums, I imagine they'll come back over the
> weekend...
>
>> What are the thoughts on this?
>>
>> Sincerely,
>>
>> Joshua D. Drake
>
> ... John
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
>
Regards,
Oleg
_____________________________________________________________
Oleg Bartunov, sci.researcher, hostmaster of AstroNet,
Sternberg Astronomical Institute, Moscow University (Russia)
Internet: oleg@sai.msu.su, http://www.sai.msu.su/~megera/
phone: +007(095)939-16-83, +007(095)939-23-83
