update string NaN values into empty strings in some tables

I realized that pandas.read_csv() by default read empty strings as NaN. Now keep_default_na=False option is added to prevent it. However, some rows are inserted as NaN, so I'd like to update these values to empty strings.