フリー問題

問題文

A data engineering team currently reads 2 TB out of Snowflake into a separate compute cluster, aggregates it there, and writes the result back. They are considering Snowpark for the same job. Which statement describes the change in where the work happens?

選択肢

1. The work is split evenly between the client process and the warehouse, so both sides have to be sized for half of the data volume.
2. The transformation is pushed down and runs inside Snowflake, so the rows no longer have to travel to the machine that runs the application code.
3. Snowpark provisions a separate cluster beside the account in the same region and runs the aggregation there, billed to the account as a distinct service that scales on its own.
4. The rows still travel to the client with transformations, but Snowpark compresses them on the way, which is where the improvement comes from.