Now import data from Snowflake into Azure for ML training

Microsoft wants you to be able to easily use data from Snowflake to train machine learning models with Azure. Importing data becomes easier, but it also somewhat negates the benefits of Snowflake.

Is your data in Snowflake? Then you can easily import it into Azure from now on. Microsoft proposes native integration of the two environments, allowing data scientists to push the right data into Azure ML with the push of a button. No additional expertise or technical effort is required to establish the connection and it is possible to import data in a structured manner according to a schedule.

Microsoft points out that companies do not store all their data in Azure. Snowflake offers what it is itself data cloud Calls. This is an environment where customers can store, share and connect all of their data, including real-time data, both internally and externally. Snowflake wears one zero copyThis means the data is in one place with all associated access policies.

Import and copy

With the import option for Azure, Microsoft seems to be ignoring this vision. If you want to use Snowflake data to train Azure AI models, you need to import that data from Snowflake into Azure ML, which in this case is a nice word to copy.

Because the data is used to train models, this isn’t a major disaster. The ease with which data scientists can access the Snowflake data is certainly an added value. Still, we can’t find a real solution Solution call up. In an ideal world, Microsoft is also thinking about an approach where data doesn’t have to be copied to different places before it can be used. Copying makes it increasingly difficult to keep control over (sensitive) data.