PostgreSQL has an "array" data type (http://www.postgresql.org/docs/9.4/static/arrays.html), as do other databases. You could use that for that purpose (I know of research prototypes that have attempted this). It is unlikely to work well in general, especially given that degree distributions are often skewed (i.e., there is very high spread in the number of relationships per node) -- if I had the time, I would do a proper empirical comparison.