RDFox ingests data in the open standard RDF-triple format which can be imported, written from scratch, or easily converted to from SQL or CSV sources.
Once triples have been imported, they are available for querying and reasoning. Additionally, triples can be scheduled for incremental addition and incremental deletion, which makes them available for incremental reasoning.
Thus, triples can be added programmatically, read from files of certain formats, or extracted from an OWL 2 DL ontology.
Converting data to RDF-triple format requires establishing a mapping pattern between the data source and the desired graph structure. The following provides an example of a mapping pattern for data in sql/csv format:
This mapping is done in a two stage process using rules:
In the following example, a company’s employee table is first converted to an employee relation and then mapped to the desired graph structure using a rule.
The materialisation of the following rule will generate the graph from the relation:
The flexibility of rules means it is possible to import data from different data sources into a the graph by first populating the binary relations in the graph as well as creating new ones.
Data can sometimes be incomplete especially when importing it from new data sources. RDFox can identify missing values or fill them in using a user defined strategy. This strategy can be applied during or after the import which improves the quality of the data.
For help converting data to RDF-triples from XML or Json, contact us at email@example.com. Find out more about RDFox or try it for yourself, for free!