Snowflake stored procedure best practices

Question

Please help me in understanding the best practices that can be followed in using stored procedures that can be called to run tasks simultaneously with tree of tasks within the root tasks called in the stored procedure.

Is this a recommended way of doing a data load? What are the efficiency impacts for such stored procedure execution?

Do share other best practices to follow for data loading through stored procedures for fact and dimension tables that I can follow apart from the above. TIA

Answer 1

Typically a TASK would call a SPROC not the other way around.

You can use TASKS to build out trees allowing for parallel execution of child tasks.

You can schedule your TASK with interval or cron, or manually execute a task with EXECUTE TASK

The data load part of your question is very broad and very specific to your use case:

• Bulk
• Streaming
• Data Source (object storage, local file system, in another DB, API, etc)
• Snowflake Native vs. External tooling (Fivetran, Matillion, HVR, etc)