From my experience with Qlik Data Integration APIs, the request-reload endpoint currently operates at the task level, not at a granular dataset level—even though the selectedDatasets parameter is present in the request body.
A couple of important points to check:
The "name" field in selectedDatasets must exactly match the internal dataset name defined in the task (case-sensitive). Even then, partial reload behavior is not consistently supported across all task types.
The "datasetId" you’re using may not be the actual internal identifier expected by the API. In most cases, this value differs from the display name and must be retrieved via metadata APIs (like listing datasets under the DI task).
As of now, Qlik DI generally does not support reloading a single dataset independently within a task via this endpoint. The API tends to trigger a full task reload regardless of dataset filtering.
Workarounds you might consider:
Splitting datasets into separate tasks if selective reload is required frequently
Using transformation logic or filters to limit data processing instead of full reloads
Checking with Qlik Support, as this behavior may vary by tenant version or may be on the roadmap
The issue is happening because the API does not recognize the dataset name when you use name, so it ends up running a full reload of all datasets. When you use datasetId, it fails because Qlik expects the internal dataset ID, not the dataset name like wd_vas_codes. You need to first get the task details, find the actual dataset ID from the response, and use that in your request. In some cases, even after using the correct ID, Qlik still reloads everything because partial dataset reload is not supported for every task type.
