Database Reference
In-Depth Information
Figure 3.30
Specifying dimensions with “Measures” hierarchy type.
For instance, if you are building an ASo model, you may want to take advantage of
multiple load buffers to expedite load process.
From a prototype perspective, it is often easier and quicker to go directly to the data
source or use a secondary flat file for the data load. From a production deployment
perspective, you should benchmark the load time results from Studio compared to
other methods to determine which method yields the most efficient and manageable
results.
3.6.2 Leveraging Open Properties to Validate Dimensionality
Everyone believes his or her data is clean. to be blunt, it never really is. his statement
is not intended to insult an organization or an individual. People honestly believe that
their data is in a good state and ready to build; there are those rare instances when that
is the case. The problem, often, is not data—it is people. Data is clean and rules are fol-
lowed until humans get involved and break the rules. over time the tightest of systems
experience a level of dirty data and metadata. Further, the definition of what represents
clean data varies from one person to another.
For instance, a client might find it perfectly acceptable that there are duplicate mem-
bers in the markets dimension when you intend to follow Essbase best practice and create
a unique member model. Inevitably the understanding and discovery of data inconsis-
tencies comes out during the load and build process. These can often be very difficult to
scrub. however, you can leverage open settings and incremental build options in Studio
to help find and resolve these inconsistencies.
