Abstract
Data completeness is commonly regarded as one of the key aspects of data quality. With this paper we make two main contributions: (i) we develop techniques to reason about the completeness of a query answer over a partially complete database, taking into account constraints that hold over the database, and (ii) we implement them by an encoding into logic programming paradigms. As constraints we consider primary and foreign keys as well as finite domain constraints. In this way we can identify more situations in which a query is complete than was possible with previous work. For each combination of constraints, we establish characterizations of the completeness reasoning and we show how to translate them into logic programs. As a proof of concept we ran our encodings against test cases that capture characteristics of a real-world scenario.