If everything (the whole DBpedia dump) is loaded to Virtuoso, it requires too much space. In order too alleviate from this burden, you can select only an essential subset of the datasets from the dump. From our experience, when you start from http://wiki.dbpedia.org/Downloads2015-10
dump, than the datasets most relevant to Odalic Semantic Table Interpreation are the following:
- ontology
- homepages
- instance types
- labels
- mappingbased literals
- mappingbased objects
- short abstracts
- wikipedia links
These are the Virtuoso iSQL commands you may use to load the datasets to the Virtuoso store:
ld_dir ('d:/KBs/DBpediaCommon', '*.owl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'homepages*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'instance_types*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'labels*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'mappingbased_literals*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'mappingbased_objects*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'short_abstracts*.ttl', 'http://dbpedia.org');
ld_dir ('d:/KBs/DBpediaEN', 'wikipedia_links*.ttl', 'http://dbpedia.org');
This must be confirmed by the iSQL command starting the import:
rdf_loader_run();