In this paper, we propose a new approach, called Query from Examples (QFE), to help non-expert database users construct SQL queries. Our approach, which is designed for users who might be unfamiliar with SQL, only requires that the user is able to determine whether a given output table is the result of his or her intended query on a given input database. To kick-start the construction of a target query Q , the user first provides a pair of inputs: a sample database D and an output table R which is the result of Q on D. As there will be many candidate queries that transform D to R , QFE winnows this collection by presenting the user with new database-result pairs that distinguish these candidates. Unlike previous approaches that use synthetic data for such pairs, QFE strives to make these distinguishing pairs as close to the original ( D,R ) pair as possible. By doing so, it seeks to minimize the effort needed by a user to determine if a new database-result pair is consistent with his or her desired query. We demonstrate the effectiveness and efficiency of our approach using real datasets from SQLShare, a cloud-based platform designed to help scientists utilize RDBMS technology for data analysis.