Querying Relational Data with Semantic Domains using SQL in Oracle RDBMS

Ontologies as well as controlled vocabularies, and domain-specific terminologies are increasingly being developed to capture semantics for life science applications. A requirement for database systems is to allow the domain of a table column to be such hierarchical terminologies and to provide query support. The paper addresses this requirement by introducing the notion of a semantic domain for relational data and providing the capability of querying over such relational data. Specifically, 1) OWL ontologies are used to model semantic domains, 2) A set of semantic-match SQL operators, namely, ONT_RELATED, ONT_DISTANCE, and ONT_PATH are introduced, to query over columns with respect to a semantic domain, 3) A new indexing scheme ONT_INDEXTYPE is introduced to speed up such queries, and 4) ONT_EXPAND operator is provided to query ontologies directly. Our approach enables users to query relational data with respect to a semantic domain directly using SQL operators, thereby opening up possibilities of combining with other operations such as joins as well as making semantically rich life science applications easy to develop and efficient. This paper describes the key concepts of querying relational data with semantic domains, discusses the set of SQL semantic match operators, and illustrates its usage via a case study involving National Cancer Institute’s Ontology.