QAV: querying entity-attribute-value metadata in a biomedical database

Entity-attribute-value (EAV) data organization is increasingly used for knowledge representation for complex heterogeneous biomedical databases. When delivered in relation form for production applications the simplicity of EAV storage is offset by difficulty of set-based data retrieval. We describe a client-server application, QAV, that is designed to perform set-based query on the Columbia MED dataset, a large medical metadata repository that has been the focus of much research. QAV interacts with the user through a graphical front end and generates a series of SQL statements that are sent to the server for the actual data retrieval.