In ParaView 5.9 when I load an XDMF dataset with a lot of parts/blocks the Blocks property tree gets truncated down to only the root like seen below.
By “a lot” I believe the cutoff is around 1000. Does anyone happen to know the actual limit? And how that limit is determined (do only leaf nodes go towards that limit or all blocks)??
I swear I’ve seen this discussed before here but I could not find that old post.
I’ve discovered that the maximum number of grids shown is set on the VTK side in vtkXdmfReaderInternal.cxx
// As soon as num-grids (sub-grids and all) grows beyond this number, we assume
// that the grids are way too numerous for the user to select individually and
// hence only the top-level grids are made accessible.
#define MAX_COLLECTABLE_NUMBER_OF_GRIDS 1000
Is there any way to override this limit, through use of environment variables or anything else??
I ask because I have my own Python reader that extends the functionality of the Xdmf2Reader. But I use the grid information to control the GridStatus in each process when running in parallel. Past this limit, only the root grid is available, so I can no longer distribute the data.
Alternatively, perhaps there’s another way to parallelize the underlying vtkXdmfReader??