Categories
Fusion Analytics Tech Tips:Semantic Model Extension Best Practices
About Semantic Model Customization
You can customize the semantic model to make the data that you moved into the warehouse more useful for reporting.
You customize your semantic model by creating branches and adding steps to those branches. You use a branch or version to publish your changes to the model. You can apply the model from the branch to a local service instance for testing. When the changes are correct, you can merge that branch with the main branch. You can merge multiple branches with the main branch over time. When you have a set of changes finalized, you can version the main branch and promote that branch to the production environment.
You can customize the model by extending ready-to-use dimensions with additional attributes from another data source, by adding a fact to an existing subject area, and by reorganizing the ready-to-use subject areas to create a new subject area to name a few.
The semantic model consists of these components:
- Oracle Content: This is the base model provided by Oracle. Your customizations are layered on this.
- System Extensions: Your descriptive flexfield and data augmentation changes are available in this component.
- User Extensions: Your customization branches and versions are available in this component.
- Security Configurations: You can secure the objects of all the other components against the application roles in this component
Attached video describes some of the best practices for Semantic Model Extension.
Comments
-
Attached is the PPT
3 -
Additional points to consider
1- Attributes added by Data Augmentation may sometimes be non-numeric - These can be converted to numeric types either from the Data Augmentation interface(recommended) , Database definition, or Semantic model (least recommended hence avoid -will have performance implications)
2- use RPD Merge feature for complex use cases ( Preview features need to open a Service Request)
3- Please take periodic backups using bundles
3 -
Some of the performance-tuning sessions can help users complete the model development more optimally
3 -
Thanks for the PTS link
2 -
Hi Abhishek, Thanks for the contents. Can you clarify comments around BP #3? Example of period rolling functions (which are notoriously impactful to performance) is given to model in the Db. However, Oracle SQL doesn't have a direct way of calc'ing time series, hence the use of
AGO
,TODATE
, andPERIODROLLING
functions in semantic model. Thus, what is the BP?1 -
Hi ..Can you check out on How to handle time-series functions
https://videohub.oracle.com/playlist/dedicated/188027173/1_crf7kgz8/1_ska8upka
1