-werners-
Esteemed Contributor III

I'd first ingest the raw data onto a data lake (using some ingest tool, databricks is not the best for this imo), then process the data using databricks.