A hybrid approach to convert relational schema to object-oriented schema

Abstract

In the paradigm of DataBase Re-Engineering (DBRE), reverse-engineering data semantics by schema translation from lower level of abstraction such as relational schema to a higher level of abstraction such as Extended Entity Relational (EER) model has been, in the past, extensively studied with due to its relative simplicity in matching. However, schema transformation from EER model to Object-Oriented DataBase (OODB) schema is not straightforward due to its missing of dynamic semantic representation, since an Object-Oriented (OO) schema should contain both the structure and operations of the data objects. In this paper, we describe a hybrid approach which applies both heuristic learning techniques in discovering the behavioural semantics from relational schema and knowledge-based approach in recovering static and structural semantics, to reach a complete conversion. A practical case is applied through an implemented prototype to validate the effectiveness of our methodology.