Azure Machine Learning datastore overview

Within an Azure Machine Learning workspace, a storage service that is the source of data is registered as a datastore for reusability. A datastore securely holds connectivity information for accessing data within the key vault that was created with your Azure Machine Learning workspace. The credentials supplied to the datastore are used to access the data within a given data service. These datastores can be created via the Azure Machine Learning Studio through the Azure Machine Learning SDK for Python, or the Azure Machine Learning command-line interface (CLI). Datastores enable data scientists to connect to data by name rather than passing connection information within scripts. This allows the portability of code through different environments (in different environments, a datastore may point to different services) and prevents the leaking of sensitive credentials.

Supported datastores include the following:

• Azure Blob Storage
...