Databases Reference
In-Depth Information
Sometimes a specific business attribute is requested but is not captured in
any current source systems. To facilitate communication, this can be included
in the dimension diagram, but the notation is different. The dotted rectangle
indicates that the data element is either not captured or is not to be included
in the initial implementation. In this example, the Holiday attribute is planned
for the future. This ensures that the model reflects true business needs, but it
also helps set and maintain expectations that the element is not going to be
available at this time.
The dimension diagram itself is only part of the documentation that is needed
for the dimension. The diagram shows each attribute with a useful business
label. In Figure 7-3, note that each of the date hierarchies is uniquely named.
This helps ensure that the correct attribute is easily selected for reporting. It is
also important to have a clear definition of each attribute, and several sample
values. The sample values are often what will spark recognition of what an
attribute represents. This does not need to be a complete list of all the possible
values for this attribute. Table 7-1 shows the additional documentation that is
needed to support the dimension diagram.
DATA NAMES AND DEFINITIONS
The business community must take a lead role in developing clear, meaningful
names for each of these attributes. The business must also be responsible for
creating the definition of each attribute. These are critical to ensure that the
model is easily understandable and accurately documented. Meaningful data
definitions are one part of overall data governance. This topic is important
enough that Chapter 8 is devoted to discussing data ownership and governance
issues.
Fact Groups
The second part of the model contains the facts, which is where the business
measurements are stored. Modeling the facts is much more than simply
creating a list of the business measures that are needed. Each of these facts
must be reflected within the proper context. This can be understood by
looking at how the data is captured and how the business uses each fact. The
dimensions that are relevant to these facts are shown. The grain, or lowest
level of detail that applies, is also identified for each applicable dimension.
Often, several facts will have the same dimensionality and identical grain.
These facts can be put together into a
fact group
.
