I do not understand why this would be a problem - are you not just saying that for a given set of data/process you are expecting a specific result (pre-deterministic testing). If this were what you are trying to do - then I would say you are following “Best Practices”.
<BLOCKQUOTE><font size="1" face="Verdana, Arial, Helvetica">quote:</font><HR>Originally posted by witchcrop: Whola!
Check out this online discussion on Data Driven Automation. This gives a good insight into how to and how not to implement Data Driven Automation testing.
Hope this will add something ...
The article you reference is interesting but totally out of date and the people involved obviously are not at the point they should be in their DDE development process. Everything they say DDE's can't do - can be done and is being done with TurboTester. Feel free to look at the site and call me for more info if interested.
<BLOCKQUOTE><font size="1" face="Verdana, Arial, Helvetica">quote:</font><HR>Originally posted by pvarakantam: I want to combine test data and expected results in data driven tests. I heard somebody saying it is not good practice to combine these two. Could any one tell me why it is not? <HR></BLOCKQUOTE>
It all depends on the situation. you have to know an expected result for your test case, so you have to have ability to get it somehow.
the good practice is the one which help you to solve problem in the most efficient way.
for example, in my automated tests:
- expected result can be provided with test data (in matrix, or encoded in the input data: e.g. specialy named files)
- expected result can be generated from the input data and not stored anywhere (actually, usually in this case both input data and expected results were generated in the run-time)
the way I choose is depends on the nuances of test.
Do what you feel is better in _your_ case. If you are sure that test data and expected result combined together will increase speed of your tests execution by eliminating unnecessary file opreations - do it.