Department of Computer, Control and Management Engineering

Incomplete information is one of the most longstanding topics in the field of data management. The notion of missing data in relational structures appears already in the early work of E. F. Codd on the relational model, and it was considered in different fields of data management ever since. Incomplete information is also a widespread phenomenon in several application scenarios. It may be the consequence, e.g., of incomplete or erroneous data insertion processes, the result of data integration, or the outcome of access restriction policies. For these reasons, the vast majority of modern database systems implement policies to perform data management tasks in the presence of incomplete information. Unfortunately, despite decades of applied and theoretical research, there is a big gap between how current systems handle incomplete information and incomplete information models studied in database theory. How big is this divide, and can we attempt to close it? In this talk, I will present some recent advancements in this direction, with a special focus on the problem of answering queries in the presence of incomplete information.