Databases Reference
In-Depth Information
Writing Analytical Applications
A typical user who works with multidimensional data goes through a process that is differ-
ent from that used by a user of relational data. From the first view of aggregated data, he
is likely to drill down and request more and more detailed data. For example, a marketing
manager of a retail enterprise might start by looking at global sales, and then focus on just
Europe. Then, he might drill down further to see the data from France, and then maybe
from just Calais, and even from one specific store. Then, he might go back to the global
view and look at just one product the company sells. And he might drill down the same
path to find out how much of that product the store in Calais sold last month or the
month before. He rarely asks for a static report; he works in constant communication with
the server.
Thus, an application for working with multidimensional data requires more flexibility and
a higher level of interactivity. The application has to connect to a multidimensional data
source, discover the structure of the stored data, and allow the user to query and drill
down into the data. Most applications intended for the analysis of multidimensional data,
even nontrivial ones, can be visualized by the diagram in Figure 33.4. You can easily write
such an application with the ADOMD.NET object model.
Make a connection to the data source
Find out which operations are supported by the
provider and prepare to use them
Find out about existing object of
metadata and their properties
Send a request for retrievals or
update of data
Close the connections
FIGURE 33.4
A client application uses this process to browse multidimensional data.
Search WWH ::
Custom Search