Possible bug? - Vector square root vs raise to power of 0.5

[Dan888]

For some reason, the square root symbol works but raising to a power of 0.5 doesn't:

Version: 15.0

Mathcad doc attached.

[MikeArmstrong]

Remove the units from a and b.

[MikeArmstrong]

Might be worth reading this:

[Werner_E]

Good information, but I am not sure if it relates to the problem here.

After all without vectorization all is working as expected.

[MikeArmstrong]

Werner Exinger wrote:

Good information, but I am not sure if it relates to the problem here.

After all without vectorization all is working as expected.

Yes but he was using column vectors, not scalars.

[Werner_E]

Mike Armstrong wrote:

Werner Exinger wrote:

Good information, but I am not sure if it relates to the problem here.

After all without vectorization all is working as expected.

Yes but he was using column vectors, not scalars.

Yes, correct. And so my conclusion is that the problem is not that of function returning different units not depending on the input units only as stated in the information you posted.

[AlanStevens]

Mathcad doesn't treat row vectors as true vectors, so best to use column vectors:

Alan

[Dan888]

Thanks, my actual vectors are quite long and look better in the worksheet as row vectors.

The square root works so I'll continue but it does seem that there is a bug in there.

Update, just played a bit more and it appears to be the decimal causing the problem. 0.5 fails but 1/2 works.

[AlanStevens]

You could just write them as row vectors, but put a transpose at the end:

e.g. a = (1 2 3)ctrl_1

Alan

[Werner_E]

Dan Rose wrote:

For some reason, the square root symbol works but raising to a power of 0.5 doesn't:

Thats a bit strange and I tend to call it a bug - at least I agree that it should work as you expected when we use column vectors.

The problem seems to be the combination of units and vectorization and the error is thrown with column and rowvectors alike.

Interestingly it works if you replace 0.5 by 1/2 and it also works with symbolic evaluation.

As Alan already noted it also works when we place the vecorization operator inside and use column vecctors.