Computer-Assisted Query Formulation

Abstract

Database management systems DBMS typically provide an applicationprogramming interface for users to issue queries using querylanguages such as SQL. Many such languages were originally designedfor business data processing applications. While these applications arestill relevant, two other classes of applications have become importantusers of data management systems: a web applications that issuequeries programmatically to the DBMS, and b data analytics involvingcomplex queries that allow data scientists to better understand theirdatasets. Unfortunately, existing query languages provided by databasemanagement systems are often far from ideal for these application domains.In this tutorial, we describe a set of technologies that assist users inspecifying database queries for different application domains. The goalof such systems is to bridge the gap between current query interfacesprovided by database management systems and the needs of differentusage scenarios that are not well served by existing query languages.We discuss the different interaction modes that such systems provideand the algorithms used to infer user queries. In particular, we focuson a new class of systems built using program synthesis techniques,and furthermore discuss opportunities in combining synthesis and othermethods used in prior systems to infer user queries.