PSQL: A query language for probabilistic relational data

摘要

Database systems based on the relational model have become very popular for the storage of large volumes of business-related data. Although the relational model provides a great range of advantages over other data models, it was designed to support deterministic data and, consequently, it lacks a comprehensive way for handling uncertain data. However, in all real-world environments, uncertainty in data values is a common occurrence. Therefore, there is a need to extend the relational model so that data uncertainty can be captured explicitly. In an earlier work, we have proposed a probabilistic relational algebra that is a consistent extension of the conventional relational algebra and is reducible to the latter. The contribution of this work is the development of a non-procedural probabilistic query language—structured around the popular SQL—that can be built on top of that algebra. We formalize the extended syntax, provide examples to illustrate the syntax, and examine query formulation from a decision-theoretic perspective.