Conjunctive Query Containment under Access Limitations

Access limitations may occur when querying data sources over the web or heterogeneous data sources presented as relational tables: this happens, for instance, in Data Exchange and Integration, Data Warehousing, and Web Information Systems. Access limitations force certain attributes to be selected in order to access the tables. It is known that evaluating a conjunctive query under such access restrictions amounts to evaluating a possibly recursive Datalog program. We address the problem of checking containment of conjunctive queries under access limitations, which is highly relevant in query optimization. Checking containment in such a setting would amount to checking containment of recursive Datalog programs of a certain class, while, for general Datalog programs, this problem is undecidable. We propose a decision procedure for query containment based on the novel notion of crayfish-chase, showing that containment can be decided in co-nexptime, which improves upon the known bound of 2exptime. Moreover, by means of a direct proof, our technique provides a new insight into the structure of the problem.