Overview
Batch-resolve up to 100 FHIRCoding inputs in a single request. The
batch never fails because one coding could not be resolved - failing
items are reported inline in the results array with an error
field, and the surrounding successful resolutions are still returned.
Use this endpoint in ETL pipelines where you are iterating over a FHIR
Bundle or exporting a FHIR dataset and need high-throughput code
resolution.
Request Body
Array of FHIR coding inputs. Each item accepts the same fields as the
single resolve endpoint:
system, code, display, vocabulary_id.
Maximum 100 items.FHIR resource type applied to every item in the batch. Used for
domain alignment checks and as the domain filter on semantic search
fallbacks. Supported values match the single resolve endpoint.
Include Phoebe recommendations on every successful resolution.
Maximum Phoebe recommendations per resolved concept (1–20).
Include a
mapping_quality signal on every successful resolution.Response Shape
Each entry inresults is either a successful resolution (same shape as
POST /fhir/resolve) or an error entry:
summary block reports totals to make monitoring easier:
| Field | Type | Description |
|---|---|---|
total | integer | Number of codings submitted |
resolved | integer | Number that resolved successfully |
failed | integer | Number that reported an inline error |
Errors
| HTTP | error.code | Cause |
|---|---|---|
| 400 | validation_error | codings missing, empty, or exceeds 100 items |
| 400 | invalid_resource_type | resource_type is not a supported FHIR resource type |
concept_not_found, unknown_system,
vocabulary_restricted) are returned inline in the results array and
do not fail the batch.