5
$\begingroup$

I am surprised that there are some cases where unit conversion works, even if the input- and output dimensions are clearly different:

UnitConvert[Quantity[1., 1/"Meters"], "Centimeters"]
(* returns 0.01/cm *)

Also, the following (and many, more complicated inputs) work. It seems that when you have a compound unit at the input, and a single unit for the output, Mathematica only converts the specified "dimension" (length, time, ...) and any power of that dimension (length^n, ...):

UnitConvert[Quantity[1., "Amperes"/"Meters"], "Centimters"]
(* returns 0.01 A/cm *)
UnitConvert[Quantity[1., "Meters"*"Kilograms"/"Amperes"], "Pounds"]
(* returns 2.20462 lb m/A *)
UnitConvert[Quantity[1., "Meters"*"Feet"*"Kilograms"/"Amperes"], "Centimeters"]
(* returns 3048 kg cm^2/A *)

While this is a convenient shortcut to simplify compound units, it also seems dangerous, since this behaviour is not specified in the documentation.

Also, I am using units in my calculations to ensure I am not making trivial errors in entering formulas or input values. I need Mathematica to throw an error when it encounters such incompatible unit conversions.

Is this a feature (to be specified as such in future versions) or a bug (likely to be fixed in future versions)? And is there perhaps a way to specify a "strict" behaviour for UnitConvert?

$\endgroup$
  • $\begingroup$ I think it is acceptable to take powers of a dimension and convert those to a compatible unit. Consider the case of an area conversion. If one had square meters and converted to centimeters, I think one would expect a result involving square centimeters and would be surprised not to get that. Negative powers should not be an exception either. Maybe others see it differently though, so far as my first claim is concerned? $\endgroup$ – Daniel Lichtblau May 1 '15 at 14:19
  • 1
    $\begingroup$ I tend to disagree: This becomes confusing for compound units such as Amperes/Meters, which you can "convert" to Centimeters. On the other hand, Mathematica throws an error when trying to convert Seconds to Megahertz, even though these are pure, reciprocral "unit-dimensions". $\endgroup$ – Martin J.H. May 4 '15 at 13:50
4
$\begingroup$

No idea concerning the interesting behaviour you mention about UnitConvert. However here is a simple workaround for you strict conversion problem (using the function CompatibleUnitQ suited for your need !!)

strictUnitConvert[q1_, q2_] := UnitConvert[q1, q2] /; CompatibleUnitQ[q1, q2];
strictUnitConvert[_, _] := "Error !! Not compatible Units";

Test

strictUnitConvert[Quantity[1., 1/"Meters"], "Centimeters"]
strictUnitConvert[Quantity[1., "Amperes"/"Meters"], "Centimeters"]
strictUnitConvert[Quantity[1., "Meters"*"Kilograms"/"Amperes"], "Pounds"]
strictUnitConvert[Quantity[1., "Meters"*"Feet"*"Kilograms"/"Amperes"], "Centimeters"]

all return

Error !! Not compatible Units

whereas for example :

strictUnitConvert[Quantity[1., "Meters"/"Seconds"], "Centimeters"/"Minutes"]

return

Quantity[6000., ("Centimeters")/("Minutes")]

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.