The logic of consistent query answering in database

点击翻译按钮获取中文摘要

For several reasons databases may contain inconsistent information, i.e. that participates in the violation of certain integrity constraints. The applicant and coauthors have characterized the consistent data as the one that is invariant under minimal forms of restoration of the consistency of the original database. The idea is to have computational access to this consistent information without having to clean the given database, and avoiding as much as possible the costly operation of computing and querying the alternative consistent versions of the database (the repairs). Rather, the approach consists in retrieving the consistent, semantically  correct information at query time. Although many progresses have been made in this area, there are still several conceptual and technical open problems in this line of research. Some of those we will address in this project are the following: (a) Developing, in semantic and operational terms, the logic that governs reasoning with and about consistent data in databases, and consistent query answering, in particular. (b) Dealing with absence of complete information in this context. (c) Possibility of composing previous (chached) consistent results in order  to consistently answer more complex queries, and the use of auxiliary views in this process. (d) Dealing with numerical attributes that have to be treated as such, as opposed to generic datatypes. (e) Consistent answering of complex queries in virtual data integration systems that easily violate global constraints. (f) Consistent query answering in non-relational database, e.g. XML, RDF. (g) Consistent exchange of data between peers, etc.