Dear ParaView Forum,
I am trying to write output from a parallel CFD solver and I am using the VTKUnstructuredGrid (.vtu) format. I wanted to save the name of the boundary patches/surfaces, but apparently there is no such mechanism to save this information in the vtu format. So, I found a (inconvenient) workaround for this in some forums. Namely, give an integer GeometryID to all cells (unique to each boundary patch) and identify boundary patches in ParaView in this way (using Threshold filter).
I have also moved on to writing the data in parallel VTU format, with each rank writing its own vtu file and the master writing also the .pvtu file. Still the boundary patches are identified using the GeometryID.
As I learned more about VTK file formats, I realized that it is possible to name each boundary patch using the vtkMultiBlockDataSet format. While I have managed to do this in serial (single .vtm file along with boundary patches in separate .vtp files), I don’t know how to do this in parallel. There appears to be no parallel vtm file format. I would appreciate if anyone can tell me if there exists a parallel vtm format or how to name/identify the boundary patches properly in the pvtu file format (instead of the GeometryID method).
While searching for this everywhere, I came to know that MultiBlockDataSet will be deprecated soon and new alternatives will be available in the future (v5.10 ~ 2021). Will this solve my problem and I just need to wait for this format?
I feel reluctant to use other CFD data formats as I like the fact that I can write vtu/vtm files directly from our code without using any libraries.