How to choose appropriate test data for equivalence class partitions?

Question

When using Equivalence Class Partitioning to generate test cases you need to choose appropriate test data (such as input values) based on your partitions.

Should I care about other test scripts generated by other means when generating my test data?

For example,

• Should I create some linkage in my documentation so that the same inputs are avoided across different scripts?
• Should I avoid common entry that would appear in use cases?
• Should I avoid boundary values that would appear in test cases generated from boundary value analysis?

Alternatively, is there an overwhelming argument of making these test scripts completely independent and therefore not worrying about the test cases in other scripts?

Answer 1

Should I create some linkage in my documentation so that the same inputs are avoided across different scripts?

Yes, you should create the same to avoid repetition of input data and expected results. Doing this your number of test cases will get reduced to required value, this will save you execution time and test case maintenance effort.

Should I avoid common entry that would appear in use cases?

As a practice of Equivalence Partitioning, your values should not overlap i.e. when partitioning is done, none of the value should be in multiple partitions or groups. Each partition will have its own set of unique values valid for that partition only.

Should I avoid boundary values that would appear in test cases generated from boundary value analysis?

You should consider these values too, as for the complete and effective testing Boundary values should also be tested. Equivalence partitioning and Boundary value analysis are closely related and can be used together at all levels of testing.

I would recommend to go through this link for a good example of Equivalence Partitioning testing.