An Informal and Efficient Approach for Obtaining Semantic Constraints Using Sample Data and Natural Language Processing

The main objective of database modelling is the design of a database that is correct and can be processed eeciently by a database management system. The eeciency and correctness of a database depends among other things on knowledge about database semantics because semantic constraints are the prerequisite for normalization and restructuring operations. Acquisition of semantic constraints remains one of the bottlenecks in database design because for most database designers formal deenition of semantic constraints is a very diicult task. Within the framework of the project RADD (Rapid Application and Database Development) experience was gathered with the informal modelling of database structures. In this paper we show an approach for the acquisition of semantic constraints which is informal , easy to understand and eecient. This method uses natural language input, sample data and a discussion of sample relations to nd out semantic constraints of a database.