Detection of Stored Procedure Bad Smells

Abstract

Stored procedures are commonly used in accessing and manipulating data in large-scale system development to optimize the query database, reduce the application workloads and reduce the traffic problems between the database and the application. If source code of stored procedures have bad smells, it will have impact in modification, and eventually have a negative impact on their quality and maintainability. This research proposes Tree Diagram and Context Analysis approach in detecting six different bad smells of stored procedures. The tree diagram approach is the comparison tree diagram of bad smells and source code which is written in PL/SQL. The context analysis approach is the creation of rules and qualifications of bad smells for increasing the accuracy in detection. In addition, this research explains the overview process, the algorithm process, and has uses example source code. The evaluation uses MI (Maintainability Index) to present the approach which are practical and effective.