SnowPro Specialty: Gen AI Question 106
Single answerTRANSLATEA global support team stores customer chat transcripts in a Snowflake table. They need a SQL-only solution to convert each transcript into English before downstream sentiment analysis, while preserving the original text in the table. The team wants the simplest approach that can be called directly in a SELECT statement without building a custom model pipeline. Which Snowflake feature is the best fit for this requirement?
- A
Use the SNOWFLAKE.CORTEX.TRANSLATE function in a query to translate each transcript to English and select both the original and translated text.
- B
Use the TRANSLATE string function to perform language translation from any source language to English directly in SQL.
- C
Use SNOWFLAKE.CORTEX.COMPLETE with a prompt asking for translation, because Cortex translation is only available through generic completion models.
- D
Export the transcripts to an external service first, because Snowflake cannot perform SQL-based language translation on stored text.
Show answer and explanation
Correct answer: A
Explanation
The best answer is to use SNOWFLAKE.CORTEX.TRANSLATE because it is the dedicated Cortex AISQL function for language translation and can be invoked directly in SQL queries against data already stored in Snowflake. This aligns with a common best practice for Gen AI workloads in Snowflake: keep processing close to the data to simplify architecture and reduce unnecessary data movement. A key exam trap is confusing Cortex TRANSLATE with the standard SQL TRANSLATE function, which only performs character substitution and is unrelated to multilingual text translation. Another plausible but less appropriate option is SNOWFLAKE.CORTEX.COMPLETE, which can handle many generative tasks but is not the most direct solution when a specialized function exists. Candidates should recognize when to use dedicated Cortex AISQL functions for practical production scenarios.
- A. Correct.
Correct. SNOWFLAKE.CORTEX.TRANSLATE is designed for translating text between languages directly in SQL. It fits the requirement for a simple SQL-only approach and can be used in a SELECT statement to return both the original transcript and an English translation for downstream processing such as sentiment analysis.
- B. Incorrect.
Incorrect. The SQL TRANSLATE string function is a traditional character-mapping function, not a natural-language translation feature. It replaces characters based on position mappings and is commonly confused with language translation because of its name.
- C. Incorrect.
Incorrect. While SNOWFLAKE.CORTEX.COMPLETE can sometimes be prompted for translation tasks, it is not the simplest or most purpose-built option when dedicated translation functionality exists. Using COMPLETE also introduces more prompt-engineering overhead and less deterministic task targeting than the dedicated TRANSLATE function.
- D. Incorrect.
Incorrect. Snowflake does support SQL-based language translation through Cortex AISQL functions. Exporting data externally adds unnecessary complexity, latency, and governance concerns when the requirement can be met directly inside Snowflake.