Preserving conceptual model semantics in the forward engineering of relational schemas

Abstract

Read online

Forward engineering relational schemas based on conceptual models (in languages such as UML and ER) is an established practice, with several automated transformation approaches discussed in the literature and implemented in production tools. These transformations must bridge the gap between the primitives offered by conceptual modeling languages on the one hand and the relational model on the other. As a result, it is often the case that some of the semantics of the source conceptual model is lost in the transformation process. In this paper, we address this problem by forward engineering additional constraints along with the transformed schema (ultimately implemented as triggers). We formulate our approach in terms of the operations of “flattening” and “lifting” of classes to make our approach largely independent of the particular transformation strategy (one table per hierarchy, one table per class, one table per concrete class, one table per leaf class, etc.). An automated transformation tool is provided that traces the cumulative consequences of the operations as they are applied throughout the transformation process. We report on tests of this tool using models published in an open model repository.

Keywords

• conceptual model semantics
• forward engineering
• relational schemas
• model-driven
• semantics preservation