Microsoft Office Tutorials and References

In Depth Information

At the end, it seems that a filter context is a perfectly good candidate for defining a set. Let

us see it with an example; start by looking at Figure 8-31, where you can see a PivotTable

working with data. You can find this example in the companion workbook CH08-09-Sets.xlsx.

FIguRE 8-31
A simple PivotTable, useful for understanding sets.

Let us analyze Figure 8-31 to try to figure out how we can define it in terms of sets. The sales

in North America in year 2002 have a value of 1,087. This value is computed by analyzing
the

set of all the sales that have been made in North America in the year 2002
. You can apply the

same technique to each cell, to discover that each cell is computed by means of analyzing a

set. Because you already know that a cell is defined by its filter context, you can think that

each cell is computed analyzing the set of values that satisfy the filter context of the cell. This

way of thinking leads you to a first definition of set, which identifies a set with a filter context.

You might be tempted to believe that a set is nothing but a filter context. Even if this were true,

we think that it is better to think of a filter context as one member of a set. So when we define

a set, we can enumerate a list of filter contexts, and the list of all those filters defines a set. We

can make an example of this using the previous figure.

We can define a set containing both
Accessories in Europe for the year 2003
and
Clothing in

North America in 2002.
Each item in the set defines a filter context, and the union of both

filter contexts (which is, in turn, a more complex filter context) is the set.

As you might have noticed, the definition of a set is given, providing values for some attributes

of the data model. You can, for example, ix the year to 2003 and define the set of all sales in

2003. In addition, you can ix another attribute—for example, the territory group—to
North

America
, and define the
sales in North America in the year 2003
. Each time you ix the value

of an attribute, you narrow the number of elements covered by the set. Whenever you do

not restrict the value of an attribute, you mean
any value is good
.