Explanation:
The solution does NOT meet the goal. While a Synapse Spark Compute can use a user-assigned identity, the proposed method of directly passing a UserAssignedIdentity class object to the SynapseSparkCompute class is incorrect. According to the community discussion and Microsoft documentation, the proper approach is to use the ManagedIdentityConfiguration class with type="UserAssigned" and pass the user_assigned_identities parameter containing UserAssignedIdentity objects. The comments with upvotes (particularly the one with 3 upvotes) show the correct code pattern: spark_compute.identity = ManagedIdentityConfiguration(type="UserAssigned", user_assigned_identities=[resource_id]). The community consensus (78% A vs 22% B) appears misleading as the detailed technical explanations support answer B.
Ultimate access to all questions.
You manage an Azure Machine Learning workspace using Python SDK v2 in Azure Machine Learning Notebooks. A Synapse Spark Compute is currently attached and uses a system-assigned identity.
You need to update the Synapse Spark Compute to use a user-assigned identity using Python code.
Proposed Solution: Pass the UserAssignedIdentity class object to the SynapseSparkCompute class.
Does the solution meet the goal?
A
Yes
B
No
No comments yet.