pass errors=None to avoid failing in the categorical case

Author: ResidentMario

pandas/core/generic.py

@@ -3443,7 +3443,7 @@ def convert_objects(self, convert_dates=True, convert_numeric=False,
 
     @Appender(_shared_docs['fillna'] % _shared_doc_kwargs)
     def fillna(self, value=None, method=None, axis=None, inplace=False,
-               limit=None, downcast=None, errors='coerce'):
+               limit=None, downcast=None, errors=None):
         inplace = validate_bool_kwarg(inplace, 'inplace')
 
         try: