Question 2
Domain 1: Design and prepare a machine learning solutionYou plan to read Parquet-format data from a folder that is not registered as a dataset using Azure ML SDK v2. What should you do?
Correct answer: B
Explanation
Azure ML SDK v2 uses data assets and job inputs to access unregistered files, and a folder of Parquet files fits a URI-based folder input. Creating a job with AssetTypes.URI_FOLDER and submitting it with ml_client.jobs.create_or_update() lets the job read the data directly from the folder path without registering a dataset.
Why each option is right or wrong
A. Create a v2 job with Dataset.File.from_files()
B. Create a v2 job with AssetTypes.URI_FOLDER and ml_client.jobs.create_or_update()
Azure ML SDK v2 treats unregistered file locations as job inputs, and a folder of Parquet files is represented as a URI-based folder input rather than a registered dataset. The job is then submitted through `ml_client.jobs.create_or_update()`, which is the v2 API for creating or updating a job definition and launching it against that folder path.
C. Create a v2 job with AssetTypes.URI_FILE and ml_client.jobs.create_or_update()