| Abstract [eng] |
Report generation is one of the most important functions when developing customer relations management systems. However, in order to display report it is needed to create a data entry form, describe SQL queries and create report templates. Because leaders want to see information about their sales, employees and clients in variety of ways it may require a lot of different reports to satisfy their needs and that in turn may require a lot of resources and time. In order to solve this problem, in was proposed to develop a method and tools to automate this tasks. It was noticed that there arises another problem after report generation tool was developed. Testing team was manually testing all reports before report generation tool was developed. But it was noticed that after developing means to automate report generation quantity of possible generated reports drastically increased and testing all possible combinations became very difficult and time consuming task. In order to solve this problem it was decided to create a cube model semantic analysis method which would test if given report cube model is correct. Correctness of model syntax is ensured by XML editing tools so it is not a concern, but even syntaxically correct model can generate invalid queries. As direct semantic model analysis would be very difficult task due to the fact that users may have some dynamic fields it would be difficult to detect related errors. So it was decided to test generated queries instead of testing model text. In order to ensure that tests can find client specific errors it was decided to use database management system as an oracle. Developed testing tool successfully detected problems in given testing model and based on statistically collected information few assumptions were made on how to improve this test. |
