Answering queries over relational databases in a privacy-preserving manner.

What is it about?

This paper presents the first solution for answering general analytical SQL queries—including Select, Project, Join, and Aggregate operations—on relational databases with foreign-key constraints while ensuring differential privacy. We further prove that this mechanism achieves optimal error for any specific input instance.

Featured Image

Why is it important?

While analytical SQL query processing over relational databases has been extensively studied, this work introduces the first solution capable of handling general analytical SQL queries under differential privacy. Additionally, we demonstrate that our solution achieves optimal error.