Column-specific missing/anomalous values

[rolyp]

Should we consider making missing/anomalous values column-specific rather than a global list? This can be similar to how it is handled in Pandas’ read_csv which is keep_default_na=False, na_values={’species’:['']}.