Query Processing in Incomplete Logical Databases

An incomplete database T combines two types of information about the real world modeled by the database : (a) the relational database with null values ("value not known") represented by axioms of a First Order Theory T0 and (b) the data dependencies that are known to be satisfied in the real world. For a given set of dependencies (functional and inclusion dependencies), a chase process transforms in two steps ("forward" and "backward" chase) type (b) information into an equivalent type (a) form. This yields a new first order theory T1 ; for a class Γ of queries (subclass of monotone queries) the evaluation on T and on T1 are equivalent. A technique involving both algebraic and theorem-proving methods provides for a sound and complete evaluation of the query.