[vtk-developers] isnan

[Michael Halle]

Goodwin Lawlor wrote:
> Michael Halle wrote:
>> I've got a dataset that uses NaN as a data value to indicate invalid 
>> pixels. A couple of places in VTK, such as vtkDataArray's GetRange() 
>> are fooled by this value.  A small test for NaN pixels would fix this 
>> problem, and native isnan() and std::isnan() don't seem to introduce 
>> significant preformance hits. (vtkDataArray could also possibly keep 
>> track of the existence of NaN values to help other classes know how to 
>> deal with them).
> 
> Yeah - vtkCurvatures could do with using NaN as a data value. How do NaN 
> values get mapped by the LUT though?


Maybe these methods on vtkScalarsToColors:


SetInvalidValue() - defaults to NaN, but could be set to something else.

SetColorForInvalidValue()

SetOpacityForInvalidValue()


If the color or opacity aren't set, behavior is undefined, I guess, and optimized filter could ignore the test for invalidity.

--Mike