But now sap have introduced a more powerful and

But now SAP have introduced a more powerful and flexible approach to modeling hierarchies
in calculation views using the graphical editor.
Since SAP HANA 2.0 SPS03, the graphical editor of the calculation view now provides a new
hierarchy function node dedicated to generating and providing analytical operations on the
generated hierarchies.
The hierarchy function node does not generate hierarchies that are to be used in BI reporting
for navigation. For that you still use the semantic node types of hierarchy. The hierarchy
functions provide plenty of rich metadata that fully describes the hierarchy in a lot of detail,
calculating useful information. You can use this rich data to further explore the hierarchy and
return answers to complex questions such as ‘How many layers of management are there
between the Head of Department and the team members (hierarchy distance) and if there is
only one level, then how many peers (hierarchy siblings) does the team member have?' This
Unit 3: Modeling Functions
© Copyright. All rights reserved. 132

cannot be achieved using the semantic node type hierarchy but this is possible with hierarchy
functions. Put simply, the hierarchy function generates rich hierarchy metadata that is meant
to be exposed and used in your analysis either projected as the final output of a calculation
view or as input to various nodes inside a calculation node.
There are two main types of hierarchy function that can be configured using this hierarchy
function node:
●
Generator — generates the basic hierarchy metadata. This is required in order to perform
functions (below) but the generated metadata can also be used without a navigation
function if only basic hierarchy information is required such as which level does an
attribute belong to, or the parent attribute for each child.
●
Navigate — performs advanced analytical operations on the generated hierarchy to
provide additional metadata that is useful for deeper analysis, such as which attributes are
the farthest away from the top level parents?
You can generate the hierarchy and then use the result as the input to the hierarchy function
in a single calculation view. In other words, in a data flow you would have two hierarchy
function nodes. The lowest one generating the hierarchy and the next one performing a
navigation function on the result.
But a better approach would be to separate the generation from the navigate function by
putting them in separate calculation views. You do this so that you are able to make changes
to the generator without affecting the navigation. Plus it means many calculation views can
reuse the same hierarchy generator result and so avoids redundancy as you will not have to
repeat the hierarchy generation in each navigation calculation view.