[Paraview] Descriptive statistics filter

[David Thompson]

Hi Martin,

>  I have an ouput file that comes from a cfd simulation. After I load
> the data in paraview and apply the descriptive statistics filter, I
> see a table presenting mean value, min., max. ... of my quantities.
> I'd like to ask if it's possible to export this table as a text file
> so that I can plot the values of certain columns with gnuplot, for
> example.

Yes, clicking on the table so it becomes the active view and then  
choosing the File->Save Data menu will let you export it as a CSV file.

>  My second question concerns the keyword 'Cardinality' in descriptive
> statistics. I'd expect this to relate somehow to the number of degrees
> of freedom of my mesh, but the numbers in 'Cardinality' column do not
> always correspond. When I make a slice of my mesh and apply the
> statistics to it, for example, the cardinality values are far bigger
> than the number of nodes of the sliced plane or curve. Is cardinality
> always the total number of nodes of the grid?

Cardinality is the number of observations encountered. For point data,  
it will be the number of nodes in the mesh (PLUS an extra entry for  
each time a node is shared on more than one process if you have  
ParaView running in parallel). For cell data, it will be the number of  
cells in mesh.

The reason that you see a large number of points for the cardinality  
of a slice is that slicing creates a new mesh whose nodes are not  
always nodes of the original mesh... they are intersections of mesh  
edges with the slice plane. If you really want properties of a subset  
of mesh nodes, you can use the Elevation filter plus a Threshold to  
select a subset of cells or points and pass that to Descriptive  
Statistics. You might also be able to use the selection mechanism to  
select nodes of the original mesh with your view's camera set to  
parallel projection (instead of perspective projection). Once you have  
the nodes selected, use the Extract Selection filter to create a new  
dataset from the selected nodes.

	Hope this helps,
	David