Currently the CSV import function only takes custom field names and when entering the custom field key it doesn't seem to recognize the field. This is prone to error since we have many fields with the same name, but they may have different field type and other distinctions. By allowing for the key to be used, it ensures accuracy.