Database Reference
In-Depth Information
specify the properties for a file, you need to delete the text file from the data source and
incrementally add it back in with the proper settings.
3.3.1.3 Think Ahead; Select a Number
Just as it is easier to think ahead and specify the
file properties correctly the first time, you also should consider the eventual creation of
the cube schema. A cube schema
requires
the existence of a numeric field in your data
source. If you do not have a numeric field, you can define one when creating the data
source. The question then arises: Why you would have a source without a numeric field?
The answer is quite simple. Depending on the nature of the data you are given, there may
or may not be a numeric field present. take for example our 39 dimension, 70-column
example. While there are columns in the text file that are most definitely relational, the
client extracted that data in a hurry and wrote the 3-gb file as all text.
Later in this chapter, I will discuss a trick for deploying a cube from Studio lever-
aging metadata only, but the most common approach is to ensure you have a numeric
field up front. When setting most file properties if you do not initially select a given
property, you can delete the file and incrementally update the data source, add the file
back, and specify the desired settings. however, this is not true for changing column
properties.
If you neglect to specify the desired data type you cannot change this within the data
source. In order to read the text file with the new column properties, you need to add a
completely new text data source, select the desired file, and then set the properties.
3.3.2 To Change Column Data Type
to change the column data type in either text or relational sources, perform the follow-
ing steps:
1. on the data source wizard, define the data source connection, click
Next
.
2. on the model tab (of the data source wizard in Figure 3.4), click
Edit Column
Properties
.
3. Click
Next
to get to the desired column, specify the data type, and click
OK
.
Note:
you also can get to the properties screen by right clicking on the column header.
3.3.3 Effectively Leveraging Relational Connections
Direct relational connections are where Studio excels. When connecting to a relational
source the breadth of available capabilities ranges from cross table joins to the full scope
of Common Platform Language (CPL is Studio's command line language functions).
From a connection perspective, however, earlier versions of Studio left quite a bit to be
desired. For example, pre-11.1.2.x versions of Studio did not allow for the incremental
updating of a relational source once it had been brought into the Studio metadata cata-
log. Instead, you had to add the new tables in a second data source. As can be seen in
Figure 3.5, the more recent versions of Studio allow for the addition of tables to a data
source via the incremental update capability.
Note:
When possible it is more stable (from a maintenance perspective) to base Studio
connections on views. Any change to column names, data types, or existence of a given
column in a table can have an impact on the hierarchical and cube schema objects based
on the data source. While view changes can have an impact as well, you (as a database
