I am refactoring a core method that has 30-odd arguments such that it will have only two.

Currently, unit tests look something like this:

[Test]
public async Task TestA1()
{

   BigTestMethod(
       paramObj: new PObj {
           attrib1 = 'a',
           attrib2 = 'b', 
           //... etc.
          attribN = 1
       },
       param1: 'something',
       param2: 'something else',
       paramObj2: new QObj {
                 qA = 1, 
                 qB = 2
     };
   }

I created a new function with 2 class parameters, and I use a TestCaseSource to set certain key values, like the below. [Variable names were changed.]

[TestCaseSource(typeof(TestSourceData), nameof(TestSourceData.SetupTestSources),
            [_pValue, null, _value, null, _money])]
public async Task TestA1(Info input, Info expected)
{
     input.attr1 = 1;
     input.attr2 = 2;
     // etc. for other attributes that are not the default values.

     expected.attr1 = 1;
     // etc.

     TestMethod_New(input, expected);
}

// ----------------
    public static class TestSourceData
    {
        private static Info _input;
        private static Info _expected;

public public static IEnumerable SetupTestSources (int p1, int p2, int p3, int p4, bool p5)
{
    // some preparatory code
    //...
    _input = new {
         param1 = p1,
         param2 = p2,
         // etc.
     }

    // same for _expected

   yield return new object[] { _input, _expected };
}

Is this a sound approach?

Would you recommend something different?