Practice
The Basics
To reinforce our understanding of the basics we have covered
so far, we will use the CrossJoin() function in a manner that
illustrates its operation in the simplest of scenarios: We will construct a simple
select query that creates an axis using CrossJoin().
As we have in many of the preceding articles of the series,
will use the MDX Sample Application as a platform from which to construct
and execute the MDX we examine, and to view the results datasets we obtain.
1.
Start the MDX
Sample Application.
2.
Clear the top
area (the Query pane) of any queries or remnants that might appear.
3.
Ensure that FoodMart
2000 is selected as the database name in the DB box of the toolbar.
4.
Select the Warehouse
cube in the Cube drop-down list box.
Let's assume for our practice example that we have received
a call from corporate Finance needing some statistics for an upcoming
presentation to the stockholders. The Finance information consumers wish to
know Warehouse Sales, Cost, and Profit balances for
operating years 1997 and 1998, by U. S. States of Store
operation.
To rephrase, the objective will be to present three annual
measures, Warehouse Sales, Cost and Profit for each of the
years 1997 and 1998, broken out by U. S. Store-State.
While we can accomplish this by cross joining the Year and State dimensions
on either the column or the row axis, the row axis will provide us a more
compact presentation. We can easily see that three measures across the columns
will result in a narrower matrix than to use the six combinations (two Years
times three U.S. States of Store operation) that we would expect
to be retrieved by an alternative arrangement of the proposed CrossJoin()
function.
Let's construct a simple query, therefore, to return the measures
as columns and the cross-joined sets as rows.
5.
Type the
following query into the Query pane:
-- MXAS18-1, Basic CrossJoin() Query
SELECT
{ [Measures].[Warehouse Sales], [Measures].[Warehouse Cost],
[Measures].[Warehouse Profit]} ON COLUMNS,
CROSSJOIN (
{[Time].[Year].Members},
{[Store].[Store Country].[USA].Children} ) ON ROWS
FROM
[Warehouse]
6.
Execute the
query by clicking the Run Query button in the toolbar.
The
Results pane is
populated by Analysis Services, and the dataset shown in Illustration 3 appears.
Illustration 3: Result Dataset - Basic CrossJoin() Example
We see the Warehouse Sales, Cost
and Profit measures populating the columns across, and the cross-product
combinations of Year (from the Time dimension, naturally) and Store-States
(from the Store dimension) appearing on the row axis.
7.
Select File
-> Save As, name the file MDX18-1,
and place it in a meaningful location.
8.
Leave the
query open for the next section.
Multiple Cross Products
Next, let's explore the use of a CrossJoin() function
in a scenario where a business need exists to cross join more than two
sets. We will say, as an illustrative example, that we have been requested to
present the measures we retrieved earlier, Warehouse Warehouse Sales, Cost
and Profit, once again for operating years 1997 and 1998,
by U. S. States of Store operation. The difference this time,
however, is that the information consumers in Finance have requested that we
present the existing row information by Store Type.
To rephrase, the objective will be to present the three
annual measures, Warehouse Sales, Cost and Profit for each
of the Store Types (there are six, including the Headquarters
classification, which appears to amount to a "non-store" group), by years
1997 and 1998, broken out by U. S. Store-State. Again, a
quick analysis of the multiples lead us to conclude that performing our cross
joins on the row axis will provide us a more compact presentation. (Six Store
Types times two Years times three U.S. States of Store
operation is 36 possible combinations, as opposed to the three columns required
by the measures, as we placed them in the columns before).
Let's construct the query to return the measures as columns
and the cross-joined sets as rows. We can begin with our initial query.
1.
Within the
query we have saved as MDX18-1, replace the top comment line of the
query with the following:
-- MXAS18-2, Nested CrossJoin() Query
2.
Save the query
as MDX18-2, to prevent damaging MDX18-1.
3.
After the
following line:
{[Measures].[Warehouse Sales], [Measures].[Warehouse Cost],
[Measures].[Warehouse Profit]} ON COLUMNS,
And before the CrossJoin() expression already in place
(shown next):
CROSSJOIN(
{[Time].[Year].Members},
{[Store].[Store Country].[USA].Children})) ON ROWS
Insert the following expression:
CROSSJOIN(
{[Store Type].[Store Type].Members},
4.
Insert a close
parenthesis ")" just before the ON ROWS keyword.
(This
serves to close the "outer" CrossJoin(), within which our
original CrossJoin() function is to be "nested.")
The
Query pane
appears as shown in Illustration 4, with the inserted coding circled in
red.
Illustration 4: The Query with Modifications Circled
NOTE: Don't forget the extra closing
parenthesis symbol ")".
5.
Execute the
query by clicking the Run Query button in the toolbar.
The
Results pane is
populated, and the dataset shown in Illustration 5 appears.
Illustration 5: Result Dataset - Nested CrossJoin() Query
6.
Re-save the
file as MDX18-2.
7.
Leave the
query open for the next section.
» 
Comment and Contribute
