Requirements
Machine Learning Extensions enhance the functionality of dab AnalyticSuite with in-database machine learning analytics directly running on your SQL Server. For this feature they use the SQL Server Machine Learning Services, which is a component that needs to be installed additionally in the SQL Server instance.
Adding Machine Learning Services to an existing SQL Server Instance
This process is documented for the following by dab AnalyticSuite supported SQL Server Versions. Besides that also refer to the dab Nexus system requirements.
Version | Documentation |
SQL Server 2017 | |
SQL Server 2019 | |
SQL Server 2022 |
To provision the required R packages one of the following methods can be used, however the first one is heavily recommended.
Using a prepackaged R runtime
To simplify the installation process we provide pre-built R versions, which already include all required binaries and packages for proper operation. They can be found in the customer portal for download. This is the recommended way to avoid issues with the package deployment and also the only way to ensure proper functionality on SQL Server 2017 and 2019.
- Download the dab:AnalyticSuite R Runtime from the customer portal
- Extract the files to a location on the SQL Server, which is accessible by the user under which the MSSQL windows service is running
- Open the command line as administrator
- Switch to the folder library\RevoScaleR\rxLibs\x64 in the extracted runtime location
- Execute the following command
.\RegisterRext.exe /configure /rhome:"%ProgramFiles%\R\R-4.2.0" /instance:"MSSQLSERVER"
Replace MSSQLSERVER with your instance name, if you have changed it and change the path for rhome to the path where you extracted the ZIP file
Installing the R runtime manually
Starting from SQL Server 2022 you can also setup a supported R installation from version 4.2 onwards, by installing the following packages using the R command for package installation:
- dplyr
- data.table
- dtringdist
- isotree
- dbscan
- stringr
- caret
- jsonlite
- ranger
- e1071
- qlcMatrix
- tidyverse
- arules
- mltools
- stopwords
