Open sylvchev opened 2 years ago
Good point, no I haven't run into this issue yet as I only use numeric extra columns. But I also have not enough experience working with hdf5 to know how to instantiate columns with arbitrary data length (if this is possible at all without defeating the purpose of hdf5?).
For now I can add the info that only fixed data size / numeric values are allowed?
Yes, it could help. As there already are columns with string values in results, I think it is possible to have variable size string.
When instancing a new evaluation that make use of additional columns capability, there could be an error if the additional columns are set to record string or non-numeric values. This results in following error inside h5py/h5t.pyx in h5py.h5t.py_create() :
TypeError: No conversion path for dtype: dtype('<U32')
It could be either documented or automatically detected. @jsosulski did you ran into a similar issue?