[Paraview] Structured Grid Blanking and Scalar Ranges

[Sean Ziegeler]

I had considered that.  The only problem is that it creates incorrect 
results in other filters.  One example is histogram (which also ignores 
blank values), where if you set the invalid values to something within 
the range you wouldn't be able to tell if it were valid or invalid 
values showing up.

I was able to add the blank array as an actual data array.  So I could 
set the values as you suggest.  And as a kludge, I can threshold the 
data set based on the blank array then do things like histogram.  Its a 
bit cumbersome going to an unstructured grid like that, but I guess it 
will have to do for now.

Thanks,
Sean

Kent Eschenberg wrote:
> I dealt with the same issue in my custom readers by setting the blanked, 
> invalid values to zero or something within the valid range.
> 
> Kent
> Pittsburgh Supercomputing Center
> 
> Berk Geveci wrote:
>> Hmmmm. Tricky. The right way of dealing with this is probably adding a
>> ComputeRange() to vtkDataSet (that takes the array) and ask it to
>> compute the range. Each sub-class could then take blanking into
>> account. This would involve quite a few changes, I am afraid.
>>
>> On 9/6/07, Sean Ziegeler <seanzig at users.sourceforge.net> wrote:
>>> I'm working on a new reader that creates a vtkStructuredGrid and sets
>>> the point visibility array for invalid values.  It works fine, except
>>> that it doesn't affect the scalar ranges.  Looking at the code, I can
>>> see that vtkPVArrayInformation does indeed not take the blanking array
>>> into account for structured grids.  It also looks like changing that
>>> behavior would be involved, as it processes each scalar array without
>>> regard to its respective containing data set, so I'm not going to worry
>>> about that just yet.
>>>
>>> Is there some other way to set the scalar range manually?
>>>
>>> Preferably in such a way that the manual range is maintained when passed
>>> through filters, but that's probably a bit much to ask.
>>>
>>> I'm working with 3.1.0 right now, but I'm willing to switch to head if
>>> necessary.
>>>
>>> Thanks,
>>> Sean
> _______________________________________________
> ParaView mailing list
> ParaView at paraview.org
> http://www.paraview.org/mailman/listinfo/paraview
>