Test effect of common_id_null mask on dataset with two or more common IDs #6
Description
Currently, the common_id_null mask is applied if any of the common ids are null. This is to give cases where one ID is missing the benefit of the doubt. However, this also means cases where one of the common_ids are null and the other is a mismatch would be given the common_id_null mask. When we have an appropriate dataset, this logic should be tested to see if it creates more false matches compared to creating a mask only for cases where all IDs are null.