There is another problem about the question with clusters. It is not very specific as to what is wanted.
Should it be a cluster where each element in the cluster is matched with a column in the database table? If so should it be matched based on cluster element name matching the column name or rather its ordinal value in the cluster match the row number? What about if the cluster doesn't match the rows in either case?
Or is it about storing a cluster as binary blob in a single column?
Questions, questions!!!!