Browse wiki

Relational databases have been designed to store high volumes of data and to provide an efficient query interface. Ontologies are geared towards capturing domain knowledge, annotations, and to offer high-level, machine-processable views of data and metadata. The complementary strengths and weaknesses of these data models motivate the research effort we present in this paper. The goal of this work is to bridge the relational and ontological worlds, in order to leverage the efficiency and scalability of relational technologies and the high-level view of data and metadata proper of ontologies. The system we designed and developed achieves: (i) automatic ontology extraction from relational data sources and (ii) automatic query translation from SPARQL to SQL. Among the others, we focus on two main applications of this novel technology: (i) ontological publishing of relational data, and (ii) automatic relational schema annotation and documentation. The system has been designed and tested against real-life scenarios from Big Science projects, which are used as running examples throughout the paper.

In this paper, we presented a completely automated approach to map relational databases and ontologies. The system proposed is capable of extracting an ontological view of the relational schema, and to enable SPARQL access to the relational data source by means of a query rewriting mechanism. The same approach can be used to efficiently store relational ontologies on a RDBMS; moreover, the mapping we devised is completely based on OWL with no need to resort to a new formalism. The impact of this system has been discussed considering three main applications: (i) publishing of relational data in an ontological format, (ii) documentation of relational schemas by means of ontological annotations, and (iii) efficient relational storage for data-intensive ontologies.