Glossary
What is FHIR Bulk Data?
FHIR Bulk Data is the FHIR specification for asynchronous, large-scale dataset export from a FHIR server, using the $export operation and NDJSON output.
In context
Bulk Data covers three scopes: system-level export ($export at the root), Group-level export ($export on a Group resource defining a cohort), and Patient-level export ($export on a single Patient). The output is one NDJSON file per resource type, suitable for loading into analytics warehouses or downstream FHIR servers.
The operation is asynchronous: the client posts the request, polls a status endpoint, and downloads the produced files when the operation completes. The pattern fits exports too large to return synchronously.
How Fire Arrow handles it
Fire Arrow supports $export at system, Group, and Patient scopes. The export resolves through the same authorization rule chain regular reads use; a research role with property filters produces a redacted export. NDJSON output is staged through configured object storage.
