Context filter problem

From: Roger Grimstad <roger_at_mediascience.no>
Date: Wed, 15 Sep 1999 16:05:38 +0200
Message-ID: <rMND3.319$uV5.2144@news1.online.no>

I'm trying to use the autorecognize filter with Oracle 7.3.3 on Solaris, but it doesn't work with word 6 or word 7.

When I try to index a Word6 document I get this in the log:

---

15:05:48 09/15/99 Begin document indexing
15:05:50 09/15/99 Errors reading documents: 0
15:05:50 09/15/99 Index data for 1 documents to be written to database
15:05:50 09/15/99    memory use: 1215132
15:05:50 09/15/99 ORA-00001: unique constraint (RTVINTRA.DR_01052_I1I1)

violated

---

And it just skips the word7 files. If I use the Word6 specific filter it works, but I also need to index word7 and html documents.

My setup:

---

begin

 ctx_ddl.create_preference ('PUB_DOCS',
                            'Docs stored in files',
                            'OSFILE');

end;

begin

     ctx_ddl.set_attribute('FORMAT','997');
     ctx_ddl.create_preference('MULTI_FILT',
                               'multiple internal filters',
                               'BLASTER FILTER');

end;

begin

     ctx_ddl.create_policy (policy_name  => 'TEZTPOL',
                         colspec   => 'CTXTEZT.TEXT',
                         textkey   => 'ID',
                         dstore_pref   => 'PUB_DOCS',
                         filter_pref   => 'MULTI_FILT',
                         wordlist_pref => 'CTXSYS.NO_SOUNDEX',
                         stoplist_pref => 'CTXSYS.NO_STOPLIST');

 end;

---

Is there a problem with my policy or is it just that the autorecognize filter is not working as it says in the documentation? Maybe getting third party filters would be a solution? I check out the Outside In filters, but they can't be used with the context engine, and that is a big disadvantage. I hope someone knows a solution to this problem!

Thanks,
Roger Grimstad Received on Wed Sep 15 1999 - 09:05:38 CDT