I don’t like this kind of statistics, and the posted question also didn’t convince me of a possible claim people applying statsitical test actually know what they do.

I did neither regard the “non-parametric” nor the “p-value” thing.

That “the distributions are different”, I would assume anyway.

To program a function returning the U-value and the z-value (where applicable following the linked advice) was rather simple.

**Errors should be expected nonetheless - as well in my code or formulas, as in guides concerning statistical tests.**

If somebody wants to actually apply the code included with the attachment, thorough considerations and checks are indispensable.

You may study the attached example. mann_whitney_U_test.ods