Chapter 4. Modeling dynamic cubes 89
7. Examine the other two dimensions.
Notice that Order Method has only one level. It might seem pointless to include Order
Method in the aggregate table, but actually it allows for a richer, more versatile aggregate
table. The set of queries that can be routed to the aggregate table is enlarged.
8. Click the Measures button.
You see two measures, Quantity and Revenue. In the mapping column, you see the
columns in the aggregate table that map to these measures. They are QUANTITY and
SALE_TOTAL.
9. Hover the mouse on one of the cells in the mapping column.
The aggregate table name and the column name appear in a tool tip. This step enables
you to trace the object mapping back to the source. For example, the tool tip for the
mapping for the measure Quantity displays the following text:
AGGR_TIME_PROD_OM_FACT.QUANTITY.
10.Return to the Aggregate cube editor and click the Level mapping tab. You see a list of
level keys, their source dimensions, and the aggregate table mapping.
11.Hover the mouse on one of the cells in the mapping column.
The aggregate table name and the column name appear in a tool tip. This step enables
you to trace the object mapping back to the source. Because
AGGR_TIME_PROD_OM_FACT is a degenerate dimension with level keys, it is sufficient
to map the cube to gosldw_sales2 with the level key mapping.
12.Return to the Aggregates tab.
13.Click the relationship edit links for any of the dimensions. The Relationship Editor opens. A
message indicates that no joins are necessary.
14.Click Cancel to close the editor.
15.Deselect the Dimension Grain check boxes for the dimension.
16.Click the Relationship edit links for that dimension.
You see that the message is gone. If you need to create a mapping to an aggregate table
through matching join keys, click the columns for the dimension and the aggregate table
and choose the keys to form a relationship. You do not need to specify level grains.
17.Click Cancel to close the Relationship Editor.
18.Click Undo to restore the level grain or manually click the level grain that was being used.
19.Click the New Dimension button. A dialog opens, which lists the dimensions that exist in
the cube but do not yet exist in the aggregate cube.
20.Click any of them and click OK. The dimension is added to the aggregate cube.
21.Select the new dimension and click Delete. The dimension is removed from the aggregate
cube.
22.Click the Slicers tab.
If you needed to add slicers, expand the member browser for the dimension, select them,
and add them to the slicer list.
23.Click the Implementation tab.
The Implementation tab shows a diagram representation of the aggregate cube. Because
AGGR_TIME_PROD_OM_FACT is a degenerate dimension, the diagram is fairly simple:
there is only one table in the aggregate cube. If the aggregate cube contained rolled-up
dimensions, they would be presented in the implementation diagram also.