Handling dirty databases: From user warning to data cleaning – Towards an interactive approach
Résumé
One can conceive many reasonable ways of characterizing how dirty a database is with respect to a set of integrity constraints (e.g., functional dependencies). However, dirtiness measures, as good as they can be, are difficult to interpret for an end-user and do not give the database administrator much hint about how to clean the base. This paper discusses these aspects and proposes some methods aimed at either helping the user or the administrator overcome the limitations of dirtiness measures when it comes to handling dirty databases.