Segmantic Model

Question

I am confused about the difference between creating a Dataset/Dataflow and building a Semantic Model. Since both can be used to create reports and dashboards, what is the critical difference between them? Could you please provide specific real-world use cases illustrating when I should use a Semantic Model instead of a simple Dataset?

Bhaskar Konar · Answer

Hi @User_PH7HJ,

Additionally, please take a look into the following posts / sessions which will be useful for the better understanding.

Fusion Analytics Extensiblity Series: Case study-Data Augmentation (Part 2 of 4) — Oracle

Fusion Analytics Extensiblity Series: Case study-Semantic Model (Part 3 of 4) — Oracle

Hope this help.

Thank you.

Bhaskar Konar · Answer

https://community.oracle.com/products/oracleanalytics/discussion/27959/segmantic-model

Hi @User_PH7HJ,

Welcome to the Oracle Analytics Community!

Please find below some high level explanation and few Docs / Links which might be useful for better understanding.

Dataflows in Oracle FDI allow users to curate data visually without coding.

Mainly used for: To extract, clean, transform, and blend data from different sources (Oracle Fusion SaaS or external apps).

Key Capabilities:

Join/Blend: Combine data from different sources (e.g., Fusion ERP + Excel file).

Transform: Add calculated columns, aggregate data, or apply machine learning models.

Automation: Can be scheduled to run to keep the dataset updated.

Semantic Model (Presentation & Governance Layer):

The Semantic Model (often called RPD in Oracle terminology) is a metadata layer that sits on top of the Autonomous Data Warehouse (ADW).

Mainly used for: To turn raw or transformed data into a business-friendly structure (Subject Areas) that ensures a 'single source of truth'.

Functionality: Defines relationships, hierarchies, and business metrics (Measures) for reporting.

Key Capabilities:

Abstracts Complexity: Hides complex database joins and SQL behind simple, named columns and measures.

Security: Enforces row-level or object-level security.

Hierarchies: Defines drill-down capabilities (e.g., Year > Quarter > Month).

https://docs.oracle.com/en/cloud/saas/analytics/23r4/fawig/data-extensions.html#GUID-C57BEDBB-394B-4C9E-89BC-10701D335362

https://docs.oracle.com/en/cloud/paas/analytics-cloud/acmdg/what-is-semantic-model.html

Hope this help.

Thank you.