Where would changes typically be placed when extending the semantic model with new attributes from another data source?

Where would changes typically be placed when extending the semantic model with new attributes from another data source?

• A. In a sandbox by editing the logical star
• B. Directly in production subject area
• C. In a separate external application
• D. In the data warehouse only

Answer: (A)

When extending a semantic model with attributes from another data source, you work in a sandbox and edit the logical star. This isolated environment lets you model the new attributes, map them to the source, and validate how they affect joins, hierarchies, and calculations without disrupting production analytics. It supports testing data quality, lineage, and impact on existing reports before any production deployment. Changing production subject areas directly risks breaking dashboards and ETL pipelines since anomalies or misalignments in the new attributes could propagate through the system. Using a separate external application would create divergence between the semantic model and how data is actually interpreted by BI tools. Limiting changes to the data warehouse only bypasses the semantic layer, leading to inconsistencies with how business users understand and interact with the data.

When extending a semantic model with attributes from another data source, you work in a sandbox and edit the logical star. This isolated environment lets you model the new attributes, map them to the source, and validate how they affect joins, hierarchies, and calculations without disrupting production analytics. It supports testing data quality, lineage, and impact on existing reports before any production deployment.

Changing production subject areas directly risks breaking dashboards and ETL pipelines since anomalies or misalignments in the new attributes could propagate through the system. Using a separate external application would create divergence between the semantic model and how data is actually interpreted by BI tools. Limiting changes to the data warehouse only bypasses the semantic layer, leading to inconsistencies with how business users understand and interact with the data.