SQL#: A Language for Maintainable and Debuggable Database Queries.

Structured Query Language (SQL) is the dominant language for managing relational databases. However, complex SQL queries are hard to write and maintain because of the intricate inter-table and inter-column relations. To this end, we propose a novel query language called SQL#, which allows programmers to construct complex queries module by module and explicitly specify the relations between different modules according to the logical steps of constructing queries. Besides, we design a SQL#-based system, aiming to facilitate the maintenance of SQL# queries. Specifically, the system renders a SQL# program into a hierarchical graph, which could help programmers understand the high-level structures of SQL# programs and the intricate relations between different components within SQL# programs. In addition, the system can ease the generation of the intermediate tables that correspond to the logical steps of constructing queries, which could help programmers debug complex SQL# queries. Notably, the design of SQL# makes it easy for the system to generate the hierarchical graph and the intermediate tables. Controlled experiments suggest that the SQL#-based system reduces the durations of writing and understanding database queries by 79% and 39%, respectively, compared to raw SQL code. [ABSTRACT FROM AUTHOR]