>>> stackme2
  State Country  group1_a1  group1_b2   Test  group2_a1  group2_b2
0    TX      US       0.45        0.3  Test1          2          6
1    MA      US       0.03        1.2  Test2          9          7
2    ON     CAN       0.70        4.2  Test3          4          2
>>>
>>> i = ['State', 'Country', 'Test']
>>> pd.wide_to_long(stackme2, stubnames=stubnames, i=i, j=j, sep='_', suffix=suffix)
                               group1  group2
State Country Test  a-b-Label
TX    US      Test1 a1           0.45       2
                    b2           0.30       6
MA    US      Test2 a1           0.03       9
                    b2           1.20       7
ON    CAN     Test3 a1           0.70       4
                    b2           4.20       2