median test or 2 sample kolmogorov test ?


I have the following problem: My data is generally non normal distributed and therefore I use the 2 sample Mann Whitney Test for comparisons.

But what shall I do if the variances (Levene Test) are not equal (I have already done transformation)?

Since now I used the Median test because this test does not make any assumptions to the data (but has also less power).

A few days ago I also realized that the 2 sample Kolmogorov Smirnoff Test exists.

What test shall I use?