COF-C03 Question 322
Single answer5.2 Explain Snowflake's data sharing capabilitiesA data provider wants to make a curated set of sales tables available to an external business partner in near real time without copying or unloading the data. The provider also wants to avoid compute charges for the partner's queries. Which Snowflake approach best meets these requirements?
- A
Create a secure share for the partner and add the required database objects to the share.
- B
Create a database replication schedule to the partner's Snowflake account so they can query a read-only copy.
- C
Unload the tables to an external stage and give the partner cloud storage access.
- D
Create a data exchange listing and require the partner to query through the provider's virtual warehouse.
Show answer and explanation
Correct answer: A
Explanation
Snowflake Secure Data Sharing is the native feature for sharing live data between Snowflake accounts without ETL, unloading, or physical copying. The provider creates a share, grants access to supported objects, and the consumer creates a database from that share. Because the consumer runs queries with its own warehouse, compute cost for querying is borne by the consumer, while storage remains with the provider. This is a core distinction from database replication, which creates replicated copies for availability and recovery use cases rather than zero-copy data sharing. Snowflake documentation on Secure Data Sharing and shares/listings emphasizes that direct shares provide live, read-only access to current data and that consumers use their own compute resources.
- A. Correct.
Correct. A secure share allows the provider to share selected database objects, such as databases, schemas, secure views, and secure UDFs, directly with another Snowflake account without copying data. Shared data is accessed in near real time because consumers query the provider's stored data through metadata pointers. For direct sharing, the consumer uses its own virtual warehouse to query the shared database, so the provider does not pay compute for the partner's queries.
- B. Incorrect.
Incorrect. Replication creates a copied database in another account or region and is designed for disaster recovery, business continuity, or cross-region/cross-cloud availability scenarios. It does not satisfy the requirement to avoid copying data. It is also not the standard mechanism for simple near-real-time partner sharing within Snowflake when direct sharing is possible.
- C. Incorrect.
Incorrect. Unloading data to external cloud storage creates copies of the data and adds operational overhead for export, file management, and refresh processes. This approach does not provide Snowflake-native near-real-time access to live shared objects and does not align with the requirement to avoid copying or unloading.
- D. Incorrect.
Incorrect. Listings in a data exchange or marketplace are used to discover and request shared datasets, but consumers still query the shared data using their own virtual warehouses after access is granted. The statement that the partner must query through the provider's warehouse is wrong. Also, if the goal is simply to share with a specific business partner, a direct secure share is the most straightforward fit.