Abstract
Functional dependencies (FDs) and inclusion dependencies (INDs) are the most important constraints in relational databases, and generalize keys and foreign keys. They are useful for many applications in database maintenance and data manipulation, bul in general they are not available in operational databases. This article presents a state of the art on IND discovery problem and gives a synthetis of our approaches for this problem. An novel data preprocessing leads to an efficient discovery of INDs of size 1. An ordered search space for INDs of size greater than I can be defined, and explored through a levelwise strategy exploiting the ami-monotonicily of IND satisfaction. We also consider the case of approximate IND discovery. We give experimental results against real-life and synthetic data.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call