SnowPro Associate: Platform Question 312
Single answer3.5 Explain how to use Snowflake Cortex LLM functions.A data engineering team stores customer support tickets in a Snowflake table and wants to generate a short summary for each ticket directly in SQL so analysts can review trends without moving data out of Snowflake. The team wants to use a built-in Snowflake Cortex large language model function in a query against the ticket text column. Which approach best meets this requirement?
- A
Use the SNOWFLAKE.CORTEX.SUMMARIZE function in a SQL statement against the ticket text column.
- B
Create a Snowpark Container Services application because Cortex LLM functions cannot be called from SQL.
- C
Use COPY INTO to export the ticket text to external storage, call a third-party LLM, and reload the results into Snowflake.
- D
Use a materialized view with automatic clustering because Cortex summaries are generated through clustering metadata rather than LLM functions.
Show answer and explanation
Correct answer: A
Explanation
This question tests whether the candidate understands the practical use of Snowflake Cortex LLM functions in SQL. For a requirement to summarize text already stored in Snowflake, the built-in Cortex SQL function for summarization is the appropriate choice. Snowflake Cortex includes SQL-accessible functions for LLM-powered tasks such as summarization, sentiment, translation, and completion-style use cases, allowing teams to process unstructured text without exporting it. Best practice is to use the native function that matches the business task and keep processing in-platform when possible. This aligns with Snowflake documentation describing Cortex AI functions as callable directly from SQL for common generative AI and text analysis workflows.
- A. Correct.
Correct. Snowflake Cortex provides built-in LLM functions that can be invoked directly from SQL for tasks such as summarization. Using SNOWFLAKE.CORTEX.SUMMARIZE on the ticket text column is the most direct way to generate concise summaries while keeping data in Snowflake.
- B. Incorrect.
Incorrect. This reflects a common misconception. Snowflake Cortex LLM functions are available directly in SQL and do not require Snowpark Container Services just to perform standard text generation or summarization tasks.
- C. Incorrect.
Incorrect. Exporting data to an external system is unnecessary for this scenario and conflicts with the requirement to avoid moving data out of Snowflake. Cortex is designed specifically to let users apply LLM capabilities within Snowflake.
- D. Incorrect.
Incorrect. Materialized views and clustering improve query performance and storage organization, but they do not perform natural language summarization. Summaries are generated by Cortex LLM functions, not by clustering metadata.