Hi,

Answering just the second question:

The triple table

CREATE TABLE BOOKSTORE_rdf_data
(id NUMBER, triple SDO_RDF_TRIPLE_S);

just holds the triples as numbers (ids)

When you create the model:

execute SEM_APIS.create_sem_model
('bookstore', 'bookstore_rdf_data', 'triple');

The view SEMM_BOOKSTORE is created

SEMM_model_name is a view with its definition based on rdf_link$ which is a partitioned table that holds all models created in the database. The SEMM_model_name is the view that represents your model in RDF_LINK$ table.

The triple table

set long 9000000
select text
from user_views where view_name = 'SEMM_BOOKSTORE';

TEXT
--------------------------------------------------------------------------------
select "P_VALUE_ID","START_NODE_ID","CANON_END_NODE_ID","END_NODE_ID","MODEL_ID"
,"COST","CTXT1","CTXT2","DISTANCE","EXPLAIN","PATH","LINK_ID" from rdf_link$ par
tition (model_3)


Hope this clarifies.

Further note. RDF_LINK$ holds ids, the real values are in RDF_VALUE$

Regards!
Jorge

Hi,

The reason for the discrepancy is that the application table allows duplicates, whereas there are no duplicate triples stored in the model.

Vladimir

*

Edited by: jbarba on Jun 14, 2010 9:59 AM

Hi Julien,

RDFB_model and RDFC_model are temporary (staging) tables used during bulk loading. They are defined as global temporary tables so when the session ends all of the data in those tables is deleted.

Vlad