QUALITATIVE ANALYSIS OF TESTING TOOLS FOR DATABASE PROCEDURES With focus on usability

Abstract: Testing is a crucial part of soft‰ware development, which is oft‰en seen as the gatekeeper of quality assurance. Testing can be divided into several di‚fferent stages of so‰ftware development. What knowledge testing requires and what tools are used in various phases is diff‚erent from each business and type of application that is under test. Th‘e functionality, tested close to the source of implementation, is almost always cheaper, easier to maintain, and less fragile both in hours and monetary costs. Code in a database is no di‚fferent from code elsewhere in the system, writing database unit tests are equally important to writing unit tests for other code. Database testing is a complicated process but many tools support developers in their work. In this report four diff‚erent tools will be compared on their usability to €find out if there is any tool that is signi€ficantly be‹tter than the others. Th‘e tools chosen are Ready API, JMeter, DBUnit, and Oracle SQL Developer, representing both open-source and commercial tools. Ready API and Oracle SQL Developer are selected by the client, Skatt‹everket. Let us look at database testing as unit tests for the code (stored procedure, views, or other complex operation) in the database. Th‘is thesis is investigating the usability of database testing tools. Considering what kind of knowledge is required from the user (e.g. tester or developer) and if the tests can be integrated with a build, test, and deploy pipeline. To understand di‚fferences in usability of tools interviews were performed among soft‰ware testers at the client, which gives a more detailed understanding of the tools. To get a broader perspective a survey within the global testing community was also performed. Based on the feedback from both the survey and the interviews there is no signi€ficant diff‚erence regarding usability for the chosen tools.