How to create hierarchy columns in Oracle Analytics Server (OAS)

User_DX76J

From this link https://youtu.be/PBxADXChCqw

I found out that it has hierarchy columns of date, but when I tried to upload dataset (or create dataset from Database connection), it does not show as the video described (not in hierarchy form)

Not hierarchy form (what I have)

Hierarchy form (what I want)

Questions

• How to create hierarchy columns in Oracle Analytics Server (OAS)?
• Do I need to use Subject Area to create hierarchy columns? (If yes, then what is the difference between normal dataset and subject area)
  • And how to create subject area? because mine shown as attached picture

SteveF-Oracle

Hi @User_DX76J ,

Please do update your display name in your profile, so the community knows with whom we are communicating with.

DV is capable to rendering hierarchies; however, they need to be defined in a subject area (Semantic Model / RPD ) for the datasource, prior to building your workbook.

Here is an example from a Local Subject Area (i.e. - dataset from a Subject Area)

The time levels when expanding a dataset date field... are not really a time hierarchy - those are just options for extracting / deriving level based columns like Year / Quarter / Month.

Creating a Subject Areas, is an entire topic on modeling

Some primers:
Time Dimension Design
Model Time Series Data

I will let other SME's chime in.

Gianni Ceresa

Steve said it all…

At this point in time, hierarchies as seen in the videos should come from the semantic model. In your case you are using a dataset, and there you don't really have a full model. It can be seen as a basic, simplistic, implicit model mostly.

In the roadmap there are hierarchies for dataset planned for the future (no date provided), not sure how it will work and what it will be.

If your data come from a database, you can model it in the semantic modeler and define the various elements (logical table, attributes, measures, logical hierarchies etc.). But as Steve said, creating a semantic model is an entire topic because you need to first get used on how it is done, how the tool works, make sure your model design is correct based on your analytics needs, build the model, deploy it and then maintain it. Also keeping in mind that the semantic model is unique in your server for all users: it's what provides the base of a "governed" analytics approach, with a unique, central, model deployed for everybody (based on permissions) to use without the users having to load data or build models.

The semantic model is extremely powerful, but depending on your needs it could be way too much work to "just" have hierarchies available…

Bhaskar Konar

Thank you very much @SteveF-Oracle & @Gianni Ceresa!

Very useful & helpful information. Thanks for sharing.