When Microsoft released R Services for SQL Server 2016 I felt a bit betrayed by Microsoft taking the other side in the data science language wars. To my good surprise they are making good with the Python community adding Python services to SQL Server 2017 (currently in preview).
In the traditional machine learning scenario we use historic data to create a model and then we use that model to predict some outcome. This presents several chalenges around the logistics of moving large chunks of data in a fast and secure way. What if we could build our models and use them to evaluate new data right where our data lives – in the SQL Server itself. And this is why Microsoft introduced the Python and R services for SQL Server.