Efficient Discovery of Functional Dependencies on Massive Data (Extended Abstract)

Xiaolong Wan, Xixian Han, Jinbao Wang, Jianzhong Li

Abstract: Functional dependencies (FDs) are the most common constraints in the design theory for relational databases, generalizing the concept of a key for a relation. Given an attribute subset $X$ and an attribute $A$ in relation schema, a functional dependency (FD) $X\rightarrow A$ means that the value of $A$ is uniquely determined by the values of $X$ . An FD $X\rightarrow A$ is minimal if $A$ is not determined by any proper subset of $X$ , and is nontrivial if $A$ is not contained in $X$ The FD discovery problem is to determine all minimal nontrivial functional dependencies over a given relation instance.