Breyta

Deila með


Run Python and R scripts in Visual Studio Code notebooks with SQL Server Machine Learning Services

Applies to: SQL Server 2017 (14.x) and later versions

Learn how to run Python and R scripts in Visual Studio Code notebooks with SQL Server Machine Learning Services. Visual Studio Code is a cross-platform development tool.

Prerequisites

Create a SQL notebook

Important

Machine Learning Services runs as part of SQL Server. Therefore, you need to use a SQL kernel and not a Python kernel.

You can use Machine Learning Services in Visual Studio Code with a SQL notebook. To create a new notebook, follow these steps:

  1. Select File and New Notebook to create a new notebook. The notebook uses the SQL kernel by default.

  2. Select Attach To and Change Connection.

    Screenshot of Visual Studio Code SQL Notebook change connection.

  3. Connect to an existing or new SQL Server. You can either:

    1. Choose an existing connection under Recent Connections or Saved Connections.

    2. Create a new connection under Connection Details. Fill out the connection details to your SQL Server and database.

    Screenshot of Visual Studio Code SQL Notebook connection details.

Run Python or R scripts

SQL Notebooks consist of code and text cells. Use code cells to run Python or R scripts through the stored procedure sp_execute_external_scripts. Use text cells to document your code in the notebook.

Run a Python script

Follow these steps to run a Python script:

  1. Select + Code to add a code cell.

    Screenshot of Visual Studio Code SQL Notebooks add code block.

  2. Enter the following script in the code cell:

    EXECUTE sp_execute_external_script
        @language = N'Python',
        @script = N'
                a = 1
                b = 2
                c = a/b
                d = a*b
                print(c, d)
                ';
    
  3. Select Run cell (the round black arrow) or press F5 to run the single cell.

    Screenshot of Visual Studio Code SQL Notebooks run Python code.

  4. The result appears under the code cell.

    Screenshot of Visual Studio Code SQL Notebook Python code output.

Run an R script

Follow these steps to run an R script:

  1. Select + Code to add a code cell.

    Screenshot of Visual Studio Code SQL Notebooks add code block.

  2. Enter the following script in the code cell:

    EXECUTE sp_execute_external_script
        @language = N'R',
        @script = N'
                a <- 1
                b <- 2
                c <- a/b
                d <- a*b
                print(c(c, d))
                ';
    
  3. Select Run cell (the round black arrow) or press F5 to run the single cell.

    Screenshot of Visual Studio Code SQL Notebooks run R code.

  4. The result appears under the code cell.

    Screenshot of Visual Studio Code SQL Notebook R code output.