A Novel Semantic Discovery Method Using Transaction Requirements in Database Reverse Engineering

Existing database reverse engineering approaches mainly analyze the database schema, especially the key constraints, to discover the relationships between relations and attributes. However these methods are only effective for direct semantics discovery, while large part of inferred semantics is still missing. This paper proposes a transaction-oriented semantic discovery method which analyzes the transaction requirements supported by a database to explore hidden semantics. Our method relaxes the constraints on normalization of database design to suit for wide range applications, and is effective for discovering transitive relationships, general-specific relationships and master-detail relationships.