Please check this example.

create table tstdel(id number, triple SDO_RDF_TRIPLE_S);
exec sem_apis.create_sem_model('tstdel','tstdel','triple');

insert into tstdel values(1, SDO_RDF_TRIPLE_S('tstdel','<urn:sub>','<urn:pred>','<urn:obj>'));

DELETE FROM tstdel s
WHERE
( to_char(s.triple.rdf_m_id, 'FMXXXXXXXXXXXXXXXX') ||'_'||
to_char(s.triple.rdf_s_id, 'FMXXXXXXXXXXXXXXXX') ||'_'||
to_char(s.triple.rdf_p_id, 'FMXXXXXXXXXXXXXXXX') ||'_'||
to_char(s.triple.rdf_c_id, 'FMXXXXXXXXXXXXXXXX') )
= SDO_RDF.GET_TRIPLE_ID('tstdel', '<urn:sub>', '<urn:pred>', '<urn:obj>');


To speed up delete, you can create a functional index on the application table.

Hi,

It is very strange. I got an email in my inbox saying that you want to find out
IDs of triples that belong to RNAIDB data set like the following.

"<http://www.lscdd.lilly.com.sg/lscdd/RNAIDB/...../.../:>".

This forum does not have your message somehow.

Assume you have asked such an question :), my answers are
1) from a modeling perspective, it is not a very good idea to encode
semantics in the URI lexical form itself. A URI should be treated
as a symbol.

2) now assume you have a valid reason for doing this, you can try something like the following.

CREATE INDEX testdel_sub_idx ON tstdel (triple.GET_SUBJECT());

-- You can then get the rowid out for those offending rows.
select rowid
from tstdel t
where t.triple.GET_SUBJECT() like '<urn:su%'
;

-- Or you can remove them directly.
delete from tstdel t
where t.triple.GET_SUBJECT() like '<urn:su%'
;

Not sure why the last line of my previous post is messed up. It should be the same as
the where clause in the "select" statement.

but the values of triple still exists in MDSYS.RDF_VALUE$ table. Is this a proper way remove triple ??

Yes, that is expected. The resource values are not removed from RDF_VALUE$ when deleting triples.

Regards,
Vladimir

Our application do atleast 1000 deletes/insert/update on triples every day.

In this case number of unused resource values in MDSYS.RDF_VALUS$ table won't affect the query performance on the model?
Or
Am i missing anything??