As a user, on the datasets, I can choose to generate iterations just having in mind the parameters mentioned on the steps

Whenever generating the iterations (due to the calculated datarows), user should be prompted if all iterations should be generated or only the necessary ones, having in mind the parameters that are referred in the steps.

Imagine these parameters paramA, paramB, paramC.

• paramA  has 2 possible values
• paramB  has 3 possible values
• paramC  has 4 possible values

This would lead to 2*3*4 = 24 iterations, even if the steps only reference paramA for example.

It that's the case, the user should have the ability to either generate all combinations... or just 2, corresponding to the 2 possible values of paramA.

Note that sometimes people may have a Test where we need to fill some parameters; meanwhile, the system evolves and then we may need just to fill one of those parameters. In this case, the current implementation would generate "useless" combinations.

Second note: some users may use datasets as contextual information, while others use them as actual data that needs to be filled somewhere in the test steps.. therefore, the suggestion for giving the user the choice to generate all combinations or just the "relevant" ones (given the actual references on the steps).