While testing the real data extraction, I encountered a new problem: the websites descriptions are rarely present in the metadata_website_desc column. Therefore, the datasets library cannot load such a dataset by having to guessing the feature type. It has to know them beforehand.
While testing the real data extraction, I encountered a new problem: the websites descriptions are rarely present in the
metadata_website_desccolumn. Therefore, the datasets library cannot load such a dataset by having to guessing the feature type. It has to know them beforehand.