The mappings described in this section assume that the cube is stored as a set of tables in an SQL schema. A precise specification of the mappings requires a correspondingly precise specification of the structure of the tables that constitute a cube. One such representation is detailed here. There are many possible ways in which a cube can be represented as a set of tables. Discussion of which particular representation is the most efficient is beyond the scope of this chapter. If the representations differ, it is straightforward to alter mappings to accommodate the differences.