Where does Azure Machine Learning Service cache data?

Question

I am looking to use Azure Machine Learning Services (the one with the new drag and drop feature; still in preview) in a new data science project.

I have realised that I can preview the data when I connect a data set; I am able to do this using the option 'Dataset output' which is available as part of the dataset.

To be able to see this data, the data needs to be cached some where.

Can someone advise where this is cached?

Answer 1

Data is cached by default in a storage account that is created along with the the ML service workspace. It has the same name as the workspace plus some numbers. Inside the account there is a blobstore called azureml-blobstore-{GUID} Inside of that container your data is cached, organized by runs.

This data is made available to ML service as a Datastore that you can navigate to in the UI by clicking "Datastores" in the blade on the left-hand of the Studio.