In old days of OBIEE 10g/11g
there was only one option to limits output of the measure column based on one
or more dimension i.e. using filters One can use constants , repository
variables, session variables, presentation variables to drive filter conditions
based on parameter selected by the users or predefined by system. The whole
purpose of applying pre-defined filter is to restrict data and allows user to
select parameters and drive report based on selected values of dimensions and
attributes.
With OBIEE 11g, introduction of
hierarchical column and selection steps makes the data limitation altogether a
different affair. There is a concept of
qualified data reference which is used to limits one or more of the dimensions
to retrieve a single value for a measure column.
This blog is an attempt to
explain Qualified Data Reference ( QDR) and how it works. The other important question around QDR is
that can session variables can be used within QDR to limit the data set to
start with and can we overwrite the session variable with prompt values.
Qualified Data Reference (QDR)
Qualified data reference (QDR)
- A QDR is a qualifier that limits one or more of the dimensions to retrieve a
single value for a measure column. Using this feature, you can qualify which
value to use for a condition evaluation. When defining a prompt, it might be
necessary to 'pin' or qualify which member value from another dimension should
be used to evaluate a condition. The QDR is the mechanism to do that.
QDR Example
The use case is that customer would like to keep only
customer types where revenue > 2,000,000 and time is prompted field.
Report Output is as shown below
QDRs are defined within the member selection step in
the Answers Analysis editor. They can be found under for section of the step
definition.
The condition step which is been defined is as follows
In short QDR is been implemented to restrict data based on
one or more dimensional reference value.
I believe you find the example I have demonstrated useful.
No comments:
Post a Comment