DAA-C01 Question 170
Single answerUse different modeling techniques for the consumption layer (e.g., dimensional, Data Vault)A retail company has consolidated point-of-sale, e-commerce, and customer loyalty data in Snowflake. The raw and integrated layers are modeled in a Data Vault 2.0 pattern to preserve source history and support traceability. Business users now need a consumption layer for Power BI dashboards that answers questions such as daily sales by product category, campaign performance by customer segment, and month-over-month store performance. Query latency and ease of use for analysts are top priorities, while the company still needs to retain the detailed historical lineage already captured upstream. Which modeling approach should the data analyst recommend for the consumption layer?
- A
Expose the existing Data Vault hubs, links, and satellites directly to Power BI because they already contain all business keys and history needed for reporting.
- B
Build a dimensional model with fact and dimension tables sourced from the Data Vault business vault or information marts, optimized for analytic queries and BI consumption.
- C
Flatten all source tables into one wide denormalized table and replace the Data Vault model entirely, because a single table is the preferred Snowflake reporting design.
- D
Create only secure views over the raw Data Vault tables and rely on BI tool joins, because semantic modeling should be handled outside Snowflake.
Show answer and explanation
Correct answer: B
Explanation
The best answer is to create a dimensional consumption layer sourced from the Data Vault foundation. In real-world modern data architectures, Data Vault is often used for integration, historization, and lineage, while dimensional models are used for analytics and dashboarding. This separation allows teams to preserve auditable source history upstream and present business-friendly, performant structures downstream. For Snowflake-based analytics, this aligns with established modeling best practices: use curated fact and dimension tables for BI workloads, especially when users need intuitive schemas and predictable query patterns. Snowflake supports either modeling approach, but for a consumption layer serving tools like Power BI, dimensional modeling is typically preferred over exposing raw Data Vault structures directly. This reflects common industry guidance around using Data Vault for enterprise data warehousing foundations and star schemas or dimensional marts for presentation and reporting.
- A. Incorrect.
Incorrect. Data Vault models are excellent for integration, auditability, and historization, but hubs, links, and satellites are not typically the best direct consumption model for BI users. They often require many joins, are less intuitive for self-service analytics, and can lead to more complex reporting logic. A common misconception is that because Data Vault stores all the data, it is automatically the best reporting model. In practice, Data Vault is usually complemented by a presentation or consumption layer such as dimensional marts.
- B. Correct.
Correct. A dimensional model with clearly defined fact and dimension tables is the standard recommendation for a consumption layer focused on performant, understandable analytics. It simplifies joins, aligns with common BI tool usage patterns, and supports measures such as sales, campaign performance, and store trends efficiently. In this scenario, the Data Vault remains the integrated historical foundation, while the dimensional layer serves downstream reporting needs.
- C. Incorrect.
Incorrect. A single wide table can work for narrow use cases, but using it as a blanket replacement for both the Data Vault and a proper consumption design is not a best practice here. It reduces flexibility, can introduce duplication and maintenance challenges, and does not preserve the separation of concerns between integrated historical storage and curated reporting structures. The misconception is that denormalization alone guarantees better analytics performance and usability.
- D. Incorrect.
Incorrect. Secure views can help with access control, but simply exposing raw Data Vault objects through views still leaves analysts and BI tools to manage complex join paths and business logic. That approach does not address the primary requirement for an easy-to-use, high-performance consumption layer. The misconception is that wrapping a complex model in views automatically makes it suitable for business consumption.