The sqlrutils package provides a mechanism for R users to put their R scripts into a T-SQL stored procedure, register that stored procedure with a database, and run the stored procedure from an R development environment.
|Built on:||R 3.4.3|
|Package distribution:||SQL Server 2017 Machine Learning Services (Windows only) and SQL Server 2016 R Services
R Client (Windows and Linux)
How to use sqlrutils
The sqlrutils library is installed as part of SQL Server Machine Learning when you add R to your installation. You get the full collection of proprietary packages plus an R distribution with its base packages and interpreters. You can use any R IDE to write R script calling functions in sqlrutils, but the script must run on a computer having SQL Server Machine Learning with R.
The workflow for using this package includes the following steps:
- Define stored procedure parameters (inputs, outputs, or both)
- Generate and register the stored procedure
- Execute the stored procedure
In an R session, load sqlrutils from the command line by typing
You can load this library on computer that does not have SQL Server (for example, on an R Client instance) if you change the compute context to SQL Server and execute the code in that compute context.
|executeStoredProcedure||Execute a SQL stored procedure.|
|getInputParameters||Get a list of input parameters to the stored procedure.|
|InputData||Input data for the stored procedure.|
|InputParameter||Input parameters for the stored procedure.|
|OutputData||Output from the stored procedure.|
|OutputParameter||Output parameters from the stored procedure.|
|registerStoredProcedure||Register the stored procedure with a database.|
|setInputDataQuery||Assign a query to an input data parameter of the stored procedure.|
|setInputParameterValue||Assign a value to the an input parameter of the stored procedure.|
|StoredProcedure||A stored procedure object.|
Add R packages to your computer by running setup for R Server or R Client:
Next, review the steps in a typical sqlrutils workflow: