ARA-C01 Question 67
Single answerObject taggingA financial services company uses Snowflake to store customer data across multiple databases and schemas. The architecture team has created a user-defined tag named DATA_CLASSIFICATION with allowed values such as PUBLIC, INTERNAL, CONFIDENTIAL, and RESTRICTED. They want to ensure that when new columns are added to business tables, classification tags are applied consistently with minimal manual effort. They also need a design that can support downstream governance reporting.
Which approach should the architect recommend?
- A
Configure automatic tag propagation so that when a tag is set on a table, Snowflake automatically assigns the same tag to all future columns created in that table.
- B
Use tag-based masking policies only, because assigning a masking policy to the tag will automatically populate the tag value on newly created columns.
- C
Apply the DATA_CLASSIFICATION tag manually or through automation at the appropriate object level, and use tag inheritance where supported so child objects can inherit the tag value from their parent object when applicable.
- D
Create separate virtual warehouses for each classification level and rely on warehouse-level tagging to classify all tables and columns queried by that warehouse.
Show answer and explanation
Correct answer: C
Explanation
The best answer is to apply object tags deliberately at the correct level and use supported inheritance behavior to reduce manual work. In Snowflake, tags are governance metadata that can be assigned to securable objects such as databases, schemas, tables, views, and columns. Architects often use automation pipelines, DDL standards, or governance workflows to apply tags consistently. Tag inheritance can help child objects reflect parent classifications in supported scenarios, but it is not a substitute for a full tagging strategy, nor does it mean every newly created column receives its own explicit tag automatically.
It is also important to distinguish object tagging from tag-based masking. Tag-based masking uses tags as a way to associate masking policies with classified columns, but the existence of a masking policy does not assign tags by itself. Likewise, tagging warehouses does not classify the data accessed through those warehouses.
For governance reporting, Snowflake provides metadata and governance visibility through account usage and information schema views related to tags and policy references, making a consistent tagging strategy valuable for audits and architecture oversight. This aligns with Snowflake best practices for data governance: define a controlled taxonomy of tags, automate assignment where possible, leverage inheritance appropriately, and use reporting views to validate coverage.
- A. Incorrect.
Incorrect. Snowflake supports tag inheritance and propagation in specific contexts, but setting a tag on a table does not mean all future columns created in that table are automatically assigned that same tag as independent column tags. A common misconception is to assume tags behave like default column properties. In practice, architects should design around supported inheritance behavior and explicit tagging processes rather than expecting automatic tagging of future columns simply because a parent table was tagged.
- B. Incorrect.
Incorrect. Tag-based masking associates masking policies with tag values so that protection can be applied consistently when tagged columns are queried. However, masking policies do not create or populate tag values on objects. This option confuses governance enforcement with metadata classification. Tags must still be assigned through DDL, automation, or inheritance where supported.
- C. Correct.
Correct. This is the most practical and accurate design. Snowflake object tagging is intended for governance metadata, and architects commonly combine explicit tagging processes with supported inheritance behavior to reduce administrative overhead. Applying the tag at the right level and leveraging inheritance where applicable helps keep classifications consistent while preserving useful metadata for governance reporting through Snowflake metadata views and governance features.
- D. Incorrect.
Incorrect. Warehouses can be tagged as Snowflake objects, but warehouse tags do not classify the data objects queried through them. Tagging compute resources is useful for cost management or operational metadata, not for applying data classification to tables or columns. This option reflects a misunderstanding between resource tagging and data-object classification.