Open fimselamse opened 7 months ago
I have also created an issue on microsoft learn: https://learn.microsoft.com/en-us/answers/questions/1532252/specifying-datastore-parameter-for-input-when-invo
Hey, thanks for the feedback. @azureml-github, can someone take a look?
Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @Azure/azure-ml-sdk @azureml-github.
Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @Azure/azure-ml-sdk @azureml-github.
@pvaneck @TajiHarrisMicrosoft Hi, any progress on this? :-)
Sorry, I'm not familiar with the codebase for this package, but the fix looks simple enough.
@azureml-github, Any reason not to propagate any user-supplied datastore name like in JobOperations
? It appears similar logic can be applied to BatchEndpointOperations
.
https://github.com/Azure/azure-sdk-for-python/blob/ba70c6a13d8f531a6783d1139120780c357450d2/sdk/ml/azure-ai-ml/azure/ai/ml/operations/_batch_endpoint_operations.py#L511C1-L515C18
The datastore parameter is never passed from the Input dict when invoking a batch endpoint with a local file. Instead it just defaults to the default datastore 'workspaceblobstorage'