Method: fhirStores.import

Full name: projects.locations.datasets.fhirStores.import

Imports resources to the FHIR store by loading data from the specified sources. This method is optimized to load large quantities of data using import semantics that ignore some FHIR store configuration options and are not suitable for all use cases. It is primarily intended to load data into an empty FHIR store that is not being used by other clients. In cases where this method is not appropriate, consider using [fhir.executeBundle][] to load data.

Every resource in the input must contain a client-supplied ID. Each resource is stored using the supplied ID regardless of the enableUpdateCreate setting on the FHIR store.

The import process does not enforce referential integrity, regardless of the disableReferentialIntegrity setting on the FHIR store. This allows the import of resources with arbitrary interdependencies without considering grouping or ordering, but if the input data contains invalid references or if some resources fail to be imported, the FHIR store might be left in a state that violates referential integrity.

The import process does not trigger Pub/Sub notification or BigQuery streaming update, regardless of how those are configured on the FHIR store.

If a resource with the specified ID already exists, the most recent version of the resource is overwritten without creating a new historical version, regardless of the disableResourceVersioning setting on the FHIR store. If transient failures occur during the import, it's possible that successfully imported resources will be overwritten more than once.

The import operation is idempotent unless the input data contains multiple valid resources with the same ID but different contents. In that case, after the import completes, the store contains exactly one resource with that ID but there is no ordering guarantee on which version of the contents it will have. The operation result counters do not count duplicate IDs as an error and count one success for each resource in the input, which might result in a success count larger than the number of resources in the FHIR store. This often occurs when importing data organized in bundles produced by [Patient-everything][fhir.Patient-everything] where each bundle contains its own copy of a resource such as Practitioner that might be referred to by many patients.

If some resources fail to import, for example due to parsing errors, successfully imported resources are not rolled back.

The location and format of the input data is specified by the parameters in ImportResourcesRequest. Note that if no format is specified, this method assumes the BUNDLE format. When using the BUNDLE format this method ignores the Bundle.type field, except that history bundles are rejected, and does not apply any of the bundle processing semantics for batch or transaction bundles. Unlike in [fhir.executeBundle][], transaction bundles are not executed as a single transaction and bundle-internal references are not rewritten. The bundle is treated as a collection of resources to be written as provided in Bundle.entry.resource, ignoring Bundle.entry.request. As an example, this allows the import of searchset bundles produced by a [FHIR search][] or [Patient-everything][fhir.Patient-everything] operation.

This method returns an Operation that can be used to track the status of the import by calling operations.get.

Immediate fatal errors appear in the error field, errors are also logged to Cloud Logging (see Viewing logs). Otherwise, when the operation finishes, a detailed response of type ImportResourcesResponse is returned in the response field. The metadata field type for this operation is OperationMetadata.

HTTP request


The URL uses gRPC Transcoding syntax.

Path parameters



The name of the FHIR store to import FHIR resources to, in the format of projects/{projectId}/locations/{locationId}/datasets/{datasetId}/fhirStores/{fhirStoreId}.

Authorization requires the following IAM permission on the specified resource name:

  • healthcare.fhirStores.import

Request body

The request body contains data with the following structure:

JSON representation
  "contentStructure": enum(ContentStructure),
  "gcsSource": {


The content structure in the source location. If not specified, the server treats the input source files as BUNDLE.



Cloud Storage source data location and import configuration.

The Healthcare Service Agent account requires the roles/storage.objectAdmin role on the Cloud Storage location.

Each Cloud Storage object should be a text file that contains the format specified in ContentStructure.

Response body

If successful, the response body contains an instance of Operation.

Authorization Scopes

Requires one of the following OAuth scopes:


For more information, see the Authentication Overview.


The content structure of the input source files.

CONTENT_STRUCTURE_UNSPECIFIED If the content structure is not specified, the default value BUNDLE is used.
BUNDLE The source file contains one or more lines of newline-delimited JSON (ndjson). Each line is a bundle that contains one or more resources. Set the bundle type to history to import resource versions.
RESOURCE The source file contains one or more lines of newline-delimited JSON (ndjson). Each line is a single resource.
BUNDLE_PRETTY The entire file is one JSON bundle. The JSON can span multiple lines.
RESOURCE_PRETTY The entire file is one JSON resource. The JSON can span multiple lines.


Specifies the configuration for importing data from Cloud Storage.

JSON representation
  "uri": string


Points to a Cloud Storage URI containing file(s) to import.

The URI must be in the following format: gs://{bucket_id}/{objectId}. The URI can include wildcards in objectId and thus identify multiple files. Supported wildcards:

  • * to match 0 or more non-separator characters
  • ** to match 0 or more characters (including separators). Must be used at the end of a path and with no other wildcards in the path. Can also be used with a file extension (such as .ndjson), which imports all files with the extension in the specified directory and its sub-directories. For example, gs://my-bucket/my-directory/**.ndjson imports all files with .ndjson extensions in my-directory/ and its sub-directories.
  • ? to match 1 character

Files matching the wildcard are expected to contain content only, no metadata.