Table A.E.5.1 - Bus tour

Please build the Aggregation Path Array and assume the following end-user requirements:

1) "a roll up by the time dimension starting at years for each bus type independent of the customer, the country the tour is going to and the employee"

2) "a monthly report on salary levels independent of the customer, the country and the bus"

3) "a complete drill-down by the time dimension down to month for each employee on a bus type, irrespective of the customer and the country"

Select the corresponding cells in the APA and choose the cubes to materialize, then highlight the derivatives of those cubes.

Figure A.E.5.1 - The resulting APA with the redundancy free-set highlighted

Size of the redundancy-free set (including the base cube): 144

Required cubes, the materialization decision and derivatives

The blue area represents the end-user requirement 1 ("a roll up by the time dimension starting at years for each bus type independent of the customer, the country the tour is going to and the employee"), requirement 2 ("a monthly report on salary levels independent of the customer, the country and the bus") and requirement 3 ("a complete drill-down by the time dimension down to month for each employee on a bus type, irrespective of the customer and the country").

Materializing cube v = (CU_*,CO_*,B_ty,E_*,T_y) (dark blue cell "T_y", representing end-user requirement 1), cube w = (CU_*,CO_*,B_*,E_sl,T_m) (dark green cell "T_m", representing end-user requirement 2) and cube x = (CU_*,CO_*, B_ty,E_em,T_m) (red cell "T_m", representing end-user requirement 3) offer us the following sets of derivatives.

Figure A.E.5.2 - Derivatives of cube v = (CU_*,CO_*,B_ty,E_*,T_y)

Figure A.E.5.3 - Derivatives of cube w = (CU_*,CO_*,B_*,E_sl,T_m)

Figure A.E.5.4 - Derivatives of cube x = (CU_*,CO_*, B_ty,E_em,T_m)

Figures A.E.5.2, A.E.5.3 and A.E.5.4 show that end-user requirement 1 and 2 are covered by the derivatives of x. Considering that v and w can easily be derived from x = (CU_*,CO_*, B_ty,E_em,T_m) by aggregation along only one dimension simplifies the materialization decision:

To meet all end-user requirements we will only have to materialize the cube x.