The problem of enabling privacy-preserving data releases has become more and more important in the last years thanks to the increasing needs of sharing and disseminating information. In this paper we address the problem of computing data releases in the form of fragments (vertical views) over a relational table, which satisfy both confidentiality and visibility constraints, expressing needs for information protection and release, respectively. We propose a modeling of constraints and of the data fragmentation problem based on Boolean formulas and Ordered Binary Decision Diagrams (OBDDs). Exploiting OBDDs, we efficiently manipulate Boolean formulas, thus easily computing data fragments that satisfy the constraints.