COF-C03 Question 278
Single answerQuery acceleration serviceA retail analytics team runs an ad hoc dashboard against a very large SALES table. The queries usually filter on a small set of product IDs and date ranges, but response times are inconsistent during peak usage. The Snowflake administrator wants to improve performance for these selective queries without simply resizing the virtual warehouse for all workloads. Which solution best addresses this requirement?
- A
Enable Query Acceleration Service on the warehouse so Snowflake can offload portions of eligible selective scan and aggregation processing to shared compute resources
- B
Create a materialized view for every dashboard filter combination because Query Acceleration Service only helps full table scans
- C
Enable Search Optimization Service because it automatically accelerates all joins, aggregations, and sorts for any query pattern
- D
Convert the warehouse to a multi-cluster warehouse because Query Acceleration Service is designed only for concurrency, not query execution speed
Show answer and explanation
Correct answer: A
Explanation
The best answer is to enable Query Acceleration Service on the warehouse. In Snowflake, QAS is designed to improve the performance of eligible queries by using shared compute resources to accelerate parts of query processing, particularly for selective workloads on large tables. This makes it a good fit when an organization wants faster response times for certain queries without increasing warehouse size for every workload. By contrast, multi-cluster warehouses are primarily for concurrency scaling, not speeding up a single query. Search Optimization Service is a different feature aimed at specific access patterns such as highly selective lookups, and it is not a blanket accelerator for all query operations. Snowflake documentation and best practices distinguish these services by use case: warehouse resizing for general compute increases, multi-cluster for concurrency, search optimization for targeted lookup access patterns, and QAS for accelerating eligible selective query workloads.
- A. Correct.
Correct. Query Acceleration Service (QAS) is intended to improve performance for eligible queries, especially selective queries that scan large tables but return relatively small result sets. It can offload parts of scan and aggregation processing to shared compute resources, helping reduce latency without requiring the administrator to permanently scale up the warehouse for all workloads.
- B. Incorrect.
Incorrect. Materialized views can help certain repeated query patterns, but creating one for every possible dashboard filter combination is impractical and not how Snowflake recommends solving broad ad hoc selectivity issues. The statement that QAS only helps full table scans is also false; QAS is specifically useful for eligible selective scans and related processing patterns.
- C. Incorrect.
Incorrect. Search Optimization Service can improve the performance of highly selective point lookup-style queries and certain join patterns, but it does not automatically accelerate all joins, aggregations, and sorts for every query pattern. This option overstates its scope and does not directly match the requirement to improve eligible selective query execution without broadly resizing compute.
- D. Incorrect.
Incorrect. Multi-cluster warehouses primarily address concurrency by adding clusters when many queries run at the same time. They do not directly accelerate the execution time of a single query in the way described here. Also, the statement about QAS being designed only for concurrency is backwards; QAS is about accelerating eligible query execution, not handling concurrent queueing.