“Why does my Azure FHIR $import job show resources as successfully imported (count > 0) but when I query the FHIR API the resources don’t exist — is this an $import commit/indexing issue?”

We are bulk importing FHIR resources (Patient, Practitioner) from Azure Data Lake Storage (ADLS) into an Azure API for FHIR workspace using the $import operation. The import job completes successfully and reports the expected counts, but when we query the FHIR service afterwards, the resources are not found (total = 0).

What We See

• Import Status Response (example)

{
  "transactionTime": "2025-09-04T10:11:32.473+00:00",
  "request": "https://<workspace>.fhir.azurehealthcareapis.com/$import",
  "output": [
    {
      "type": "Practitioner",
      "count": 10,
      "inputUrl": "https://<storageaccount>.blob.core.windows.net/adls-to-fhir/care_manager_fhir/care_managers_batch_1.ndjson"
    }
  ],
  "error": []
}

Job completes with correct counts (Imported Practitioner resources: 10).

When validating via GET {{FHIR_URL}}/Practitioner?_count=10 → response shows:

{ "resourceType": "Bundle", "total": 0, "entry": [] }

Direct read via GET {{FHIR_URL}}/Practitioner/{id} also returns 404 Not Found.

Additional Info

Import configuration in workspace:

"importConfiguration": {
  "enabled": true,
  "initialImportMode": false,
  "integrationDataStore": "<storageaccount>"
}

Using UUIDs for every resource id (auto-generated, always unique).

Import manifest example:

{
  "resourceType": "Parameters",
  "parameter": [
    { "name": "inputFormat", "valueString": "application/fhir+ndjson" },
    { "name": "input", "part": [
        { "name": "type", "valueString": "Practitioner" },
        { "name": "url", "valueUri": "https://<storageaccount>.blob.core.windows.net/adls-to-fhir/care_manager_fhir/care_managers_batch_1.ndjson" }
      ]
    }
  ]
}

If we POST a single Practitioner/Patient via the FHIR REST API, it works fine and the resource is retrievable.

Error logs container (import-error-logs) is empty.

Expected Behavior

After $import reports success with count > 0, resources should be retrievable via FHIR REST API (GET /Patient, GET /Practitioner).

Actual Behavior

Resources are not retrievable (total=0 or 404) even though import shows success.

Question

Has anyone else faced this issue where $import reports success but resources are not accessible afterwards? Is this related to the $import commit/indexing behavior, and how can we reliably ensure imported resources are available?We are bulk importing FHIR resources (Patient, Practitioner) from Azure Data Lake Storage (ADLS) into an Azure API for FHIR workspace using the $import operation. The import job completes successfully and reports the expected counts, but when we query the FHIR service afterwards, the resources are not found (total = 0).

What We See

Import Status Response (example)

{

Job completes with correct counts (Imported Practitioner resources: 10).

When validating via GET {{FHIR_URL}}/Practitioner?_count=10 → response shows:

{

Direct read via GET {{FHIR_URL}}/Practitioner/{id} also returns 404 Not Found.

Additional Info

Import configuration in workspace:

"importConfiguration":

Using UUIDs for every resource id (auto-generated, always unique).

Import manifest example:

{

If we POST a single Practitioner/Patient via the FHIR REST API, it works fine and the resource is retrievable.

Error logs container (import-error-logs) is empty.

Expected Behavior

After $import reports success with count > 0, resources should be retrievable via FHIR REST API (GET /Patient, GET /Practitioner).

Actual Behavior

Resources are not retrievable (total=0 or 404) even though import shows success.

Question

Has anyone else faced this issue where $import reports success but resources are not accessible afterwards? Is this related to the $import commit/indexing behavior, and how can we reliably ensure imported resources are available?