Re: Intermedia URL_DATASTORE problem: Huge index never ending

Problem solved:

Nobody responded but we found the solution:

By default, the URL_DATASTORE does NOT use the INSO filter, contrary to the FILE_DATASTORE that does. The result was that each and every word (as well as postcript/pdf code, html tags, MS-Word control characters, etc) were indexed, which resulted in an un unmanageable and HUGE intermedia index.

"Guy Dallaire" <gd-newsgroups_at_spamex.com> a écrit dans le message de news:QkWKb.41524$BA6.903197_at_news20.bellglobal.com...
> I have a table containg 7000 local file names that I index using
intermedia
> using a FILE_DATASTORE. Indexing takes a while, and eventually, I get an
> index ($I table of about 240 Mb)
>
> Now, the application server and the database server will reside on
different
> machines, and the files are no longer local to the database server and we
> decided to index using an URL_DATASTORE. The file content is the same.
The
> DBMS and App server are on the same local network.
>
> The problem is that now, the index take forever to load (it never actually
> completes, I have to kill it) and is HUGE. When I killed it, the $I table
> was 2.4 Gb !
>
> Also, we noticed that the $I (Token table) file, we have lots of junk
(Weird
> symbols, etc...) that do not exist when we use the FILE_DATASTORE.
>
> I will open a TAR but wanted to know if anyone experienced the same
problem.
>
> Thanks
>
>
Received on Thu Jan 08 2004 - 08:19:29 CST