Test structure advice for webservices
I wondered if anyone could share their experience of best practice when approaching web service testing.
I am new to QTP, having had 1 weeks training and am now ready to go! I have automated two web service tests that are quite straightforward, and now need to do something more difficult.
The test cycle is to generate a unique ref number for a transaction using a web service, then use that ref to enter a transaction using a different web service, and then query the transaction status using a 3rd web service. All the data exists in a spreadsheet so it is easily data driven.
What I am not sure of is how best to structure the tests.
I have two options: Do all the generating calls first, storing the ref numbers with the transactional data. Then do all the data entry, and finally do all the queries at the end.
Secondly, genereate the ref, enter trans and query for each transaction and then do the next one etc.
I am not sure what would be the best way to approach this so any ideas are welcome.
On a related point, the generate ref webservice is one I have automated, using 5 simple tests. If I want to call this reusable action from a new test, can I replace the data used on the original action with different data without affecting the original test? I presume I just add the data I want to use to the Global sheet in the new test and make sure that the correct data sheet is referenced in the Value field - is that correct?.
Excuse my ignorance, but as I say, I am new to this! I would rather ask the experts than waste days trying to work out what to do. Time is money as they say!!
Steve W [img]/images/graemlins/cool.gif[/img]