Database Reference
In-Depth Information
that I did not state that it treats them in the same manner, it does not. Studio provides a
much more robust set of capabilities when leveraging relational sources. however, when
first exploring data during prototyping for a deployment, it is important to simply get
the data. to that end, it is just as advantageous to get access to data in a flat file as it is in
a relational connection.
3.3.1 You Do Not Need 39 Dimensions: Working with Text Sources
It is worth repeating, Studio does not provide the same breath of capability when work-
ing with text sources as it does when working with relational. For example, when work-
ing with text sources, you do not have the ability to create joins and, thereby, create
dimensions by combining multiple text files. The only exception is performing a recur-
sive join on a text file to enable a parent/child build. nonetheless, Studio does provide a
robust means of both introspecting and modeling with text.
A question often asked is whether or not it is worth using Studio if you are leveraging
flat files? As an experienced Essbase developer, there is no doubt that you are adept at
working with Load rules (whether in prototyping or production). however, depending
on the nature of the file and the dimensional requirements of the business, Studio can be
significantly easier. This is easily illustrated by telling a story about the 39 dimensions.
“We would like 39 dimensions,” the client said with great confidence. Their current
BI deployment had 39 dimensions, so they felt that they Essbase model should have
39 dimensions as well. It is important to note, at this point that they were talking about
base dimensions, not attributes. They were convinced that Essbase was the proper solu-
tion, but they also were convinced that they needed to see what 39 dimensions with
their data would look like. The sample source data was provided in one 3-gb file with
70 columns as seen in table 3.1.
Looking at the names of the columns in the file, it is obvious to a true online ana-
lytical processing (oLAP) person that much of this information is either not relevant to
the solution or can be leveraged in other ways (such as with an attribute or uDA [user-
defined attribute]). The client, being newer to Essbase had to be convinced. In this situa-
tion, you can either sit through multiple meetings explaining the dimensional options or
you can build the prototype the client is asking for along with a prototype of your own
design. In leveraging Studio, our team was able to build the 39 dimensions in a single
day from the test file and a “more proper” prototype in another day. to have done the
same thing with the Data Prep Editor would have taken the better part of a week (bar-
ring major error). to have retrofitted the file into a relational model would have involved
three days of It (Information technology) requests and another few days of setting of
the system. Studio's flat file functionality allowed for quick conversations and modeling
that cut through the clutter and came up with a good business solution for long-term
deployment. In short, you should absolutely consider leveraging Studio for flat files.
3.3.1.1 Importance of Server.Properties
While it may seem remedial, it is important to
remember that all text sources for Studio need to reside in a single root directory. This
directory is determined by the Studio server.properties file. This master directory can
contain subdirectories, and it need not be on the same physical volume as Studio. It sim-
ply denotes a consistent location for data file location (not dissimilar from a system DSn
(data source name) pointing to and setting properties for a relational source). Within
the file, the specific property line denoting the data file location is
server.datafile.dir
as
shown in Figure 3.1.
