Databases Reference
In-Depth Information
Data warehouse schemas are typically arrived at through the dimensional modeling
approach. The business processes of interest are determined. For example, a company
may be interested in exploring data from scheduling, productivity tracking, and job
costing. The data from each business process is destined to become a star schema. Some
of the business processes may share dimensions. For example, the Cost_Center dimen-
sion is meaningful for scheduling, productivity tracking, and job costing. A useful tool
for capturing the commonality of dimensions between business processes is the data
warehouse bus. Table 14.1 shows a data warehouse bus where each row represents a
business process and each column represents a dimension. Each x indicates that the
Table 14.1
Example Data Warehouse Bus
given dimension is applicable to the business process. The dimensions that are shared
across business processes should be “conformed.” That is, each dimension and its levels
should be known by the same names and have the same meanings across the entire
enterprise; likewise for the values contained by said dimension levels. This is important
so that data can be compared across business processes where meaningful, and people
can discuss the data in the same terms from department to department, facilitating
meaningful communication. The dimension data can be thought of as flowing through
the data warehouse bus.
The data warehouse schema can contain multiple star schemas, with shared dimen-
sions as indicated by the data warehouse bus. Figure 14.8 illustrates the data warehouse
schema corresponding to the data warehouse bus shown in Table 14.1. The attributes
are elided since we are focusing on the fact tables and dimensions. We have marked the
fact tables using a «fact table» stereotype, whereas the tables that are not marked are
dimension tables. The configuration of multiple star schemas in a data warehouse forms
a constellation.
