As highlighted by #187, there are files containing index for certain data sources, but with empty datasets.
Currently reading a selection of data sources containing one of these would fail reading data.
Should we skip these sources when datasets have exactly zero entries instead of raising an exception?
We could also catch the exception for any method which returns data from multiple datasets and exclude that key (as if it had a valid index but no data).
As highlighted by #187, there are files containing index for certain data sources, but with empty datasets.
Currently reading a selection of data sources containing one of these would fail reading data. Should we skip these sources when datasets have exactly zero entries instead of raising an exception?