Query Reverse Engineering

Given today's ease of collecting data and the need for ad-hoc data querying to explore the data, various data analysis techniques are being developed to help data consumers gain a better understanding of their data. This project investigates a useful data analysis technique based on the concept of query reverse engineering. In conventional database querying, the user poses a query Q on a database D to obtain a result Q(D). In query reverse engineering, the problem is reversed in that the inputs are a database D and a query result R, and the goal is to derive a query Q (or multiple queries) such that R is the result of query Q on database D.

Query reverse engineering can be applied to derive alternative interpretations of query results [SIGMOD 2009, VLDB Journal 2014] and to gain better insights into query results by explaining the presence/absence of certain records in the query results or certain relationships among records [SIGMOD 2010].

Selected Publications

• Q.T. Tran, C.Y. Chan, S. Parthasarathy, Query Reverse Engineering, The VLDB Journal, Vol. 23, No. 5, 2014, 721-746.
• Q.T. Tran, C.Y. Chan, How to ConQueR Why-not Questions, Proceedings of ACM SIGMOD Intl' Conference, Indianapolis, Indiana, June 2010, 15-26.
• Q.T. Tran, C.Y. Chan, S. Parthasarathy, Query by Output, Proceedings of ACM SIGMOD Intl' Conference, Providence, Rhode Island, June 2009, 535-548.

Acknowledgements