What determines the number of data files processed in parallel during a load operation?

• A. The amount of compute resources in a warehouse
• B. The total size of all files
• C. The network bandwidth between storage and compute
• D. The number of data files to be loaded

Answer: (A)

Parallelism during a load is driven by the warehouse’s compute resources. The load operation uses multiple worker processes to parse and ingest files, and how many of those workers can run at once depends on how much compute capacity the warehouse has. More compute resources mean more parallel loaders working on different files at the same time, so you can process more files concurrently.

The total size of the files mainly affects how long the load takes, not the number of files that can be processed in parallel. Network bandwidth is typically not the limiting factor inside Snowflake’s managed service, and while having many files can provide opportunities for parallelism, the actual degree of parallelism is set by the warehouse size.