Semantic model vs custom dataset joins for manager headcount

Question

Hi everyone,

I’m building an Oracle Analytics dashboard and hitting limitations with dataset joins.

Setup:

* Dataset A: Recognition data (Office, Team, Compensation Manager) — includes some terminated employees

* Dataset B: Unique award recipients (same grain)

* Dataset C: Active headcount (Office, Team, Compensation Manager)

Requirement:

Show, by Compensation Manager,

recognized colleagues, total active headcount

Problem:

When joining on Office + Team + Compensation Manager, headcount drops because Dataset A doesn’t have all manager combinations / team codes / office codes. Adding dummy rows would blow up the data.

Questions:

1. Is the Semantic Model RPD the better approach for this multi-fact scenario?

2. Any documentation or step-by-step guide to bring custom datasets into the semantic model and implement this correctly?

(Note: I used AI to help refine this problem statement for clarity.)

Thanks in advance!

Silvio Garber-Oracle · Answer

Have you tried changing the join type? 
Understand Dataset Table Joins

JohnW-Oracle · Answer

Hello @Jestin Rajan,

This question may be a little too involved for the community. Awards, management hierarchies, terminations, etc are not implemented the same way by every organization.

In general, we suggest Data Augmentations plus Semantic Model Extensions. If you have a solution that works in OTBI or using SQL, then that could help you to get started with the solution in FDI. The fact that you understand the 'grain' differences helps. I assume a left outer join between dataset A and C did not product the correct results. I am not sure what you meant by 'Adding dummy rows would blow up the data.' but the FDI documentation does have reporting and filter recommendations for each Subject Area. With Terminations or Manager Headcount, sometimes adding the Termination Date and/or the Effective Start and End dates (even as hidden fields) can help.

Regards,

John