ADA-C01 Question 478
Single answerConfigure and manage Snowflake Data Clean RoomsA media company is deploying Snowflake Data Clean Rooms to let an advertising partner measure campaign overlap against the media company's audience data. The media company must ensure that: (1) the partner cannot directly view raw audience records, (2) only approved analysis logic can be executed, and (3) the clean room can be safely reused with multiple partners over time. Which action should the Snowflake administrator take to best meet these requirements?
- A
Create a secure data share of the audience tables directly to the partner and require the partner to query only through secure views.
- B
Configure a Snowflake Data Clean Room, link only the approved datasets, and publish only administrator-approved analysis templates so collaborators can run controlled queries without direct access to base data.
- C
Replicate the audience data into the partner's account and protect it with dynamic data masking policies and row access policies managed by the partner.
- D
Create an application package that exposes SQL worksheets to the partner so they can write their own overlap queries while warehouse access is restricted.
Show answer and explanation
Correct answer: B
Explanation
The best answer is to use Snowflake Data Clean Rooms as intended: configure the clean room, link only the data that should participate in collaboration, and expose only approved analysis templates or governed analysis logic. This enables privacy-enhancing collaboration without directly sharing raw records with a partner. It also supports repeatable onboarding of multiple collaborators under a controlled framework.
This scenario tests the administrator's ability to distinguish a clean room implementation from more general Snowflake sharing or policy-based controls. Secure shares, secure views, masking policies, and row access policies are important Snowflake governance features, but they do not replace the clean-room-specific model of controlled collaboration and approved analyses. In practice, Snowflake Data Clean Rooms are used when organizations need to collaborate on sensitive data while limiting exposure and tightly governing what can be computed.
Best-practice reasoning aligns with Snowflake documentation for Data Clean Rooms: use clean rooms for governed multi-party analysis, link permitted datasets, and publish only approved templates or analyses rather than allowing unrestricted SQL access to underlying data.
- A. Incorrect.
Incorrect. A secure share can expose data objects such as tables and secure views, but it does not by itself provide the governance model required for Data Clean Rooms, including controlled collaboration workflows and restricting collaborators to only approved analysis logic. Even if secure views are used, this approach relies too heavily on what the partner can query and is not the best fit for clean-room-style governed joint analysis.
- B. Correct.
Correct. Snowflake Data Clean Rooms are designed for privacy-enhancing collaboration between parties. Administrators can configure the clean room, link the datasets that are permitted for collaboration, and control which templates or analysis logic collaborators are allowed to run. This supports the stated requirements: raw records are not directly exposed, only approved analysis patterns are available, and the clean room model is reusable across collaborations.
- C. Incorrect.
Incorrect. Replicating the data into the partner's account significantly increases exposure risk and weakens governance because the partner would control the environment where the copied data resides. Masking and row access policies are valuable Snowflake features, but this option does not align with the clean room objective of avoiding direct raw-data access while enforcing centrally approved analysis workflows.
- D. Incorrect.
Incorrect. Application packaging is not the primary control model for Snowflake Data Clean Rooms, and allowing the partner to write arbitrary SQL worksheets conflicts with the requirement that only approved analysis logic be executed. Restricting warehouse access does not solve the problem of governing query behavior against protected collaborative datasets.