Relational database design using dependency-extended entity-relationship model

Abstract

Entity-relationship model is widely used as a database design tool because the model uses diagrammatic techniques and makes possible formal modeling of real world without considering the data structures of the underlying database management system. However, conventional entity-relationship models are not sufficient to describe the mapping constraints of complex relationships and inclusion constraints between relationships. In this thesis, a dependency-extended entity-relationship model is proposed, which is sufficient to describe all mapping constraints and inclusion constraints. Complex relationships are decomposed and/or projected according to the mapping dependencies imposed on the relationships. In order to design entity-relationship schema interactively, a new incremental approach is proposed. In the incremental approach, the information described in the ER model is classified by its types, and the entity-relationship schema is incrementally constructed by the way of appending each type of information to the current schema interactively. A translation algorithm for generating the relational schema from an entity-relationship schema is also proposed. The translation algorithm preserves the functional dependencies and inclusion constraints, and the relation schemes generated by the algorithm are in the third normal form.