問題文
A data scientist wants to run a Python training routine that reads a table, fits a model, and writes the artifact back, all inside Snowflake and on a schedule. Which construct is the right container for that routine?
選択肢
- A scalar user-defined function, because a function receives the rows it needs and can return the fitted model itself as the value of its single output column.
- An external function, with a task that runs on a schedule, because the training has to happen outside Snowflake for the artifact to be written back.
- A Python stored procedure, because it can execute a sequence of operations including reads and writes and can be called by a task.
- A view, because a view can hold the training query and will be evaluated on the warehouse that the account uses whenever the schedule fires.