Problem with linterp in Mathcad 10.0.1

[Bob_Bijlsma]

Hello,

In Mathcad Prime 9.0.0.0 I am able to calculate the value of the program. In Mathcad Prime 10.0.1 I get the following error message:

What is happening? The two attached calculation sheets are identical. 

When I define each step in the program seperately in Mathcad Prime 10.0.1, it does give me the right answer.

Also, when I replace all the infinity symbols by say 10000, the problem still arises.

Hope you can help me!

Kind regards

[ttokoro]

in table29 makes the error.

[Werner_E]

I can confirm that your P9 worksheet returns a result in Prime 9 but the very same worksheet throws the error in Prime 10.

I would suggest that you contact official PTC support.

As @ttokoro  found out, the text "z/x" in the top left corner of your table causes the problem. If you replace it by any number or by a NaN, the sheet works in Prime 10 as well.

I guess its a bug in Prime 10, as this text is effectively removed by the various submatrix expressions is is never in play during the calculation.

So the program should work OK in Prime 10 as well but it doesn't. You should file a bug report.

[Werner_E]

I am pretty sure now that its a very strange bug!

I looked further into it and finally found the a bug in reversing your vector x.list introduces a NaN out of the blue instead of the correct value 0. This NaN then the causes the error when this vector is used in "interpolate" (first column of Matrix A.:

Needless to say that in Prime 9 the output is correct (a 0 and not a NaN).

As a workaround so you are not stopped at your work you can replace the text "z/x" by a NaN, but this bug sure should be reported.
I am not sure how to do this best - it may be especially difficult if you don't have access to your company's  sales order number etc.

Maybe @DJNewman  could be of help once again...

[Werner_E · Accepted solution]

Additional note:

I tracked down the bug to this simple example:

Obviously every zero is replaced by a NaN.

Both, "submatrix" and "reverse" have to be applied for the bug to show in Prime 10.

The attached worksheet is in Prime 9 format, so you can test easily that it works OK in Prime 9, but fails after recalculation when loaded in Prime 10.

Prime 9:

EDIT: Even simpler without programming and transposing