SQL Server Machine Learning and Programming Extensions Documentation

Learn how to use R and Python external libraries and languages on resident, relational data with our quickstarts, tutorials, and how-to articles. R and Python libraries in SQL Server machine learning include base distributions, data science models, machine learning algorithms, and functions for conducting high-performance analytics at scale, without having to transfer data across the network.

In SQL Server 2019, Java code execution uses the same extensibility framework as R and Python, but does not include data science and machine learning function libraries. For more information about new features, see What's New in SQL Server Machine Services.

R logo Open-source R, extended with RevoScaleR and Microsoft AI algorithms in MicrosoftML. These libraries give you forecasting and prediction models, statistical analysis, visualization, and data manipulation at scale.
R integration starts in SQL Server 2016 and is also in SQL Server 2017.
Python logo Python developers can use Microsoft revoscalepy and microsoftml libraries for predictive analytics and machine learning at scale. Anaconda and Python 3.5-compatible libraries are the baseline distribution.
Python integration starts in SQL Server 2017.
Java logo Java developers can use the Java language extension to wrap code in stored procedures or in a binary format accessible through Transact-SQL.
Java integration starts in SQL Server 2019 - preview.

SQL Server Machine Learning R and Python Documentation

Learn how to use R and Python external libraries and languages on resident, relational data with our quickstarts, tutorials, and how-to articles. R and Python libraries in SQL Server machine learning include base distributions, data science models, machine learning algorithms, and functions for conducting high-performance analytics at scale, without having to transfer data across the network.

R logo Open-source R, extended with RevoScaleR and Microsoft AI algorithms in MicrosoftML. These libraries give you forecasting and prediction models, statistical analysis, visualization, and data manipulation at scale.
R integration starts in SQL Server 2016 and is also in SQL Server 2017.
Python logo Python developers can use Microsoft revoscalepy and microsoftml libraries for predictive analytics and machine learning at scale. Anaconda and Python 3.5-compatible libraries are the baseline distribution.
Python integration starts in SQL Server 2017.

5-Minute Quickstarts

Step-by-Step Tutorials

Video Introduction

Reference

Package Language Description
RevoScaleR R Distributed and parallel processing for R tasks: data transformation, exploration, visualization, statistical and predictive analytics.
MicrosoftML R Functions based on Microsoft's AI algorithms, adapted for R.
olapR R Imports data from OLAP cube.s
sqlRUtils R Helper functions for encapsulating R and T-SQL.
revoscalepy Python Distributed and parallel processing for Python tasks: data transformation, exploration, visualization, statistical and predictive analytics.
microsoftml Python Functions based on Microsoft's AI algorithms, adapted for Python.