Databases Reference
In-Depth Information
Figure 5.7
Product graph in Calendar and Customer dimensions.
reduced disk I/O and faster response. The Profit_by_Invoice_Date (i.e., the “date, all”
node) is the best choice, requiring the reading of only one block.
5.4
Resource Considerations
There are a number of resource constraints to keep in mind as you design your material-
ized views. These include the number of materialized views, the extra disk space
required, and the length of the available update window. We discuss each in turn.
The number of possible views in a typical database is tremendous. Given a set of
dimensions, the number of views is the product of the number of levels in each dimen-
sion. The number of possible views for the example in Figure 5.7 is 4 Calendar levels
×
5 Customer levels
2 Job levels = 40 possible views. Remember, we left out a level of
the Job dimension in order to make the graph readable. We have only examined the
×
