It is not clear what is the benefit of nodes without edges approach other than needing to map the results. In most cases you will need to map the dataset either way, and all this does is adds multiple paths to access the same data (which is not necessary a good thing).