Revision as of 11:11, 4 May 2015

Downloads

The packages are available for download at sourceforge.net

LUBM-SP: http://sourceforge.net/projects/graphke/files/lubmsp_1.0.0.jar/download

LUBM-SP ontology: http://sourceforge.net/projects/graphke/files/lubmsp_onto.nt/download

SPARQL queries http://sourceforge.net/projects/graphke/files/queries.tar.gz/download

Jena Query evaluation http://sourceforge.net/projects/graphke/files/lubmsp_query_eval_1.0.0.jar/download

Instructions

Setting up files

1. mkdir lubmsp
2. cd lubmsp
3. Download all the files provided in the Download section above into this folder
4. tar -xzvf queries.tar.gz
5. You may want to install the rlwrap package for command history

Here we explain the step-by-step to generate the LUBM-SP datasets, and run the query evaluation

Generating LUBM and LUBM-SP datasets

1. cd lubmsp
2. Run the code, the number of university may be 50, 500, or 1000
java -jar lubmsp_1.0.0.jar -univ 500 -index 0 -onto http://knoesis.org/lubmsp
3. Check for three data files generated
LUBM_500.nt: the original LUBM in the form of ntriples
LUBM_500_sp.nt: the LUBM-SP in the form of ntriples
LUBM_500.rdf: the original LUBM in the form of RDF/XML

Query evaluation in Jena

1. cd lubmsp
2. java -jar lubmsp_query_eval_1.0.0.jar 
or
rlwrap java -jar lubmsp_query_eval_1.0.0.jar
reasoner:
rdfs
ontology:
lubmsp_onto.nt
instances:
LUBM_500_sp.nt
query:
queries/lubm_query1.sparql

Preparing Oracle 12c

1. Download and install Oracle database 12c to $ORACLE_HOME
2. Download and install Oracle client: sqlplus
3. Create one Oracle instance using dbca 
4. Connect to sqlplus
sqlplus / as sysdba
4. Execute the two files $ORACLE_HOME/md/admin/catsem.sql and $ORACLE_HOME/md/admin/sdordfwm.sql
SQL> @'?/md/admin/catsem.sql'
SQL> @'?/md/admin/sdordfwm.sql'
5. Create a tablespace rdf_tblspace
6. Create a semantic network on the tablespace rdf_tblspace

Below is the sample code for loading LUBM_500_sp.nt to Oracle

connect user/pass

set timing on

create table stable_lubm_500 ( RDF$STC_sub varchar2(4000) not null, RDF$STC_pred varchar2(4000) not null, RDF$STC_obj varchar2(4000) not null);

grant SELECT on stable_lubm_500 to MDSYS;

CREATE TABLE RDF$ET_TAB_LUBM_500 ( proc_sid VARCHAR2(30), proc_sig VARCHAR2(200), event_name varchar2(200), start_time timestamp, end_time timestamp, start_comment varchar2(1000) DEFAULT NULL, end_comment varchar2(1000) DEFAULT NULL );

GRANT INSERT on RDF$ET_TAB_LUBM_500 to MDSYS;

host sqlldr userid=vinh/knoesis12345 control=/semweb1/lubmsp/lubm1000/bulkload_lubm_500.ctl data=/semweb1/lubmsp/lubm1000/LUBM_500.nt_aa direct=true rows=10000000 skip_index_maintenance=true skip=0 load=1000000000 discardmax=1000000 bad=d0.bad discard=d0.rej log=d0.log errors=1000000

host sqlldr userid=vinh/knoesis12345 control=/semweb1/lubmsp/lubm1000/bulkload_lubm_500.ctl data=/semweb1/lubmsp/lubmsp_onto.nt direct=true skip=0 load=1000000000 discardmax=1000000 bad=d0.bad discard=d0.rej log=d0.log errors=1000000

commit;

create table atab_lubm_500 (triple sdo_rdf_triple_s);

exec sem_apis.create_sem_model('LUBM_500_MODEL', 'ATAB_LUBM_500', 'TRIPLE');

grant INSERT,UPDATE on atab_lubm_500 to MDSYS;

exec sem_apis.bulk_load_from_staging_table(model_name=>'LUBM_500_MODEL', table_owner=>'vinh', table_name=>'stable_lubm_500', flags=>'PARSE PARALLEL_CREATE_INDEX parallel=6');

commit;

select count(*) from table(sem_match('{?s ?p ?o}', SEM_MODELS('LUBM_500_MODEL'), NULL, NULL, NULL));


INSERT INTO ATAB_LUBM_500 VALUES (
SDO_RDF_TRIPLE_S('LUBM_500_MODEL', 
'<http://www.w3.org/1999/02/22-rdf-syntax-ns#singletonPropertyOf>', 
'<http://www.w3.org/2000/01/rdf-schema#subPropertyOf>', 
'<http://www.w3.org/2000/01/rdf-schema#subPropertyOf>'));

commit;

EXEC SEM_APIS.CREATE_ENTAILMENT('rdfs_rix_lubm_500', SEM_Models('LUBM_500_MODEL'), SEM_Rulebases('RDFS'));

commit;

Loading data to Oracle 12c

There are different ways to load data into Oracle. Here we use the bulk load from staging table method.