Hi there, wondering what the Snowflake integration is going to include? Was looking into solutions like Google Health's [FHIR data pipes](https://developers.google.com/open-health-stack/fhir-analytics/data-pipes) to transform FHIR data into SQL-formatted data, but hopeful that the Medplum <> Snowflake integration may handle some of that logic!

Hey @Maxine Whitely, good question! We definitely get a good number of requests for a Snowflake integration or some other data lake/warehouse solution. We are not 100% committed to Snowflake though. Ideally, it will provide a way to bulk export and flatten FHIR resources. This is a feature that we hope to build, but it is not actively being worked on. If you are interested in how we have seen other people build these connectors, a common pattern we have seen is to first ingest FHIR resources into a data warehouse as raw JSON, and then flatten the relevant fields in a second ETL stage. Analytics pipelines can use the [FHIR Bulk Data Export API](https://www.medplum.com/docs/api/fhir/operations/bulk-fhir) to extract FHIR resources into a data lake. The [FHIR Datastore](https://www.medplum.com/docs/fhir-datastore) supports bulk export. If you need to de-identify or redact data for compliance reasons if you are going to be using this data for analytics, you can use AccessPolicies to do that.