so the other datastore implementations have unique indices that prevent duplicate relationships from being created, this is something that we have to do manually for memdb