Hacker News Reader

I have 'one' exception to that. I worked with a medical data analysis company that had implemented SAP HANA as a data warehouse/analysis platform for tons of treatment records.

My task was replacing it with Spark/S3 :P

View on HN

My current workplace has implemented a custom SAP HANA-based solution for forecasting (code is written in a mixture of Python and R). I wasn't involved in the project but I'm not convinced it couldn't have been architected completely differently - perhaps a suite of microservices running on K8S with a PostgreSQL backend. Having an in-memory database licensed by RAM usage quickly gets mother-f'ing expensive as Nathan Fielder would say.

Reply on news.ycombinator.com