Currently I am working on assignment where need to test lots of BI reports, and for validating these reports i have been working on SQL queries which can provide the output and compare it with the BI published Report created by the dev team.
I am finding the above approach of writing SQL's and validating the reports good but at the same time challenging as well as SQL's are too long and in case if even a single condition is missed it will lead to different set of results.
Client is raising the question on deadlines and suggesting to test the reports using the same object variables which are exposed under different analytics. But i have a reservation against that and have raised the query that "If i will also start using the same variables then how different will be my analysis/qa then there existing reports?”
So would like to get some more feedback from the experts regarding how to proceed further for testing of BI report, without using OATS or any other tool initially.
Currently as per my BI Strategy following is the testing i am doing -
1- Unit testing (although primarily responsibility of dev team)
2- Functional testing (testing report or individual repository object/variable)
3- Integration testing (testing individual variable with different dimension under test)
Pref testing/security testing are other candidate but currently keeping them out of scope.
Kindly help me gurus/seniors based on your experience/knowledge. Thanks in advance for your time and effort.