Abstract
NoSQL applications often use denormalized databases in order to meet performance goals, but this introduces complications as the database itself has no understanding of application-level denormalization. In this paper, we describe a procedure for reconstructing a normalized conceptual model from a denormalized NoSQL database. The procedure’s input includes functional and inclusion dependencies, which may be mined from the NoSQL database. Exposing a conceptual model provides application developers with information that can be used to guide application and database evolution.
Type
Publication
International Conference on Conceptual Modeling
Michael Mior
Assistant Professor
My research focuses on data integration and understanding for non-relational data.