Cmdlets for Databricks

Build 24.0.9060

Establishing a Connection

With the CData Cmdlets users can install a data module, set the connection properties, and start scripting. This section provides examples of using our Databricks Cmdlets with native PowerShell cmdlets, like the CSV import and export cmdlets.

Installing and Connecting

If you have PSGet, installing the cmdlets can be accomplished from the PowerShell Gallery with the following command. You can also obtain a setup from the CData site. 

Install-Module DatabricksCmdlets

The following line is then added to your profile, loading the cmdlets on the next session: 

Import-Module DatabricksCmdlets;

You can then use the Connect-Databricks cmdlet to create a connection object that can be passed to other cmdlets: 

$conn = Connect-Databricks -Server "127.0.0.1" -HTTPPath "MyHTTPPath"" -User "MyUser" -Token "MyToken"

Connecting to Databricks

To connect to a Databricks cluster, set the following properties:

  • Database: The name of the Databricks database.
  • Server: The Server Hostname of your Databricks cluster.
  • HTTPPath: The HTTP Path of your Databricks cluster.
  • Token: Your personal access token. You can obtain this value by navigating to the User Settings page of your Databricks instance and selecting the Access Tokens tab.

You can find the required values in your Databricks instance by navigating to Clusters and selecting the desired cluster, and selecting the JDBC/ODBC tab under Advanced Options.

Configuring Cloud Storage

The cmdlet supports DBFS, Azure Blob Storage, and AWS S3 for uploading CSV files.

DBFS Cloud Storage

To use DBFS for cloud storage, set the CloudStorageType property to DBFS.

Azure Blob Storage

Set the following properties:

  • CloudStorageType: Azure Blob storage.
  • StoreTableInCloud: True to store tables in cloud storage when creating a new table.
  • AzureStorageAccount: The name of your Azure storage account.
  • AzureAccessKey: The storage key associated with your Databricks account. Find this via the azure portal (using the root account). Select your storage account and click Access Keys to find this value.
  • AzureBlobContainer: Set to the name of you Azure Blob storage container.

AWS S3 Storage

Set the following properties:

  • CloudStorageType: AWS S3.
  • StoreTableInCloud: True to store tables in cloud storage when creating a new table.
  • AWSAccessKey: The AWS account access key. You can acquire this value from your AWS security credentials page.
  • AWSSecretKey: Your AWS account secret key. You can acquire this value from your AWS security credentials page.
  • AWSS3Bucket: The name of your AWS S3 bucket.
  • AWSRegion: The hosting region for your Amazon Web Services. You can obtain the AWS Region value by navigating to the Buckets List page of your Amazon S3 service, for example, us-east-1.

Authenticating to Databricks

CData supports the following authentication schemes:

  • Basic
  • Personal Access Token
  • Azure Active Directory (AD)
  • Azure Service Principal

Basic

Basic authentication requires a username and password. Set the following:

  • AuthScheme: Basic.
  • User: Your username. This overrides the default value ("Token").
  • Token: Your password.

Personal Access Token

To authenticate, set the following:

  • AuthScheme: PersonalAccessToken.
  • Token: The token used to access the Databricks server. It can be obtained by navigating to the User Settings page of your Databricks instance and selecting the Access Tokens tab.

Azure Active Directory

To authenticate, follow these steps:

  1. Register an application with the AzureAD (now known as Microsoft Entra ID) endpoint in the Azure portal. See Configure an app in Azure portal for information on how to create and register the application. Alternatively, you can use a AzureAD application that is already registered.

  2. Set these properties:

    • AuthScheme: AzureAD.
    • AzureTenant: The "Directory(tenant) ID" in the AzureAD application "Overview" page
    • OAuthClientId: The "Application(client) ID" in the AzureAD application "Overview" page.
    • CallbackURL: The "Redirect URIs" in AzureAD application "Authentication" page

  3. When connecting, a web page opens that prompts you to authenticate. After successful authentication, the connection is established.

Here is an example of the connection string: 

"Server=https://adb-8439982502599436.16.azuredatabricks.net;HTTPPath=sql/protocolv1/o/8439982502599436/0810-011933-odsz4s3r;database=default;
AuthScheme=AzureAD;InitiateOAuth=GETANDREFRESH;AzureTenant=94be69e7-edb4-4fda-ab12-95bfc22b232f;OAuthClientId=f544a825-9b69-43d9-bec2-3e99727a1669;CallbackURL=http://localhost;"

Azure AD Service Principal

To authenticate, set the following properties:

  • AuthScheme: AzureServicePrincipal.
  • AzureTenantId: The tenant ID of your Microsoft Azure Active Directory.
  • AzureClientId: The application (client) ID of your Microsoft Azure Active Directory application.
  • AzureClientSecret: The application (client) secret of your Microsoft Azure Active Directory application.

Retrieving Data

The Select-Databricks cmdlet provides a native PowerShell interface for retrieving data: 

$results = Select-Databricks -Connection $conn -Table "[CData].[Sample].Customers" -Columns @("City, CompanyName") -Where "Country='US'"
The Invoke-Databricks cmdlet provides an SQL interface. This cmdlet can be used to execute an SQL query via the Query parameter.

Piping Cmdlet Output

The cmdlets return row objects to the pipeline one row at a time. The following line exports results to a CSV file: 

Select-Databricks -Connection $conn -Table [CData].[Sample].Customers -Where "Country = 'US'" | Select -Property * -ExcludeProperty Connection,Table,Columns | Export-Csv -Path c:\my[CData].[Sample].CustomersData.csv -NoTypeInformation

You will notice that we piped the results from Select-Databricks into a Select-Object cmdlet and excluded some properties before piping them into an Export-CSV cmdlet. We do this because the CData Cmdlets append Connection, Table, and Columns information onto each row object in the result set, and we do not necessarily want that information in our CSV file.

However, this makes it easy to pipe the output of one cmdlet to another. The following is an example of converting a result set to JSON: 

 
PS C:\> $conn  = Connect-Databricks -Server "127.0.0.1" -HTTPPath "MyHTTPPath"" -User "MyUser" -Token "MyToken"
PS C:\> $row = Select-Databricks -Connection $conn -Table "[CData].[Sample].Customers" -Columns (City, CompanyName) -Where "Country = 'US'" | select -first 1
PS C:\> $row | ConvertTo-Json
{
  "Connection":  {

  },
  "Table":  "[CData].[Sample].Customers",
  "Columns":  [

  ],
  "City":  "MyCity",
  "CompanyName":  "MyCompanyName"
}

Deleting Data

The following line deletes any records that match the criteria: 

Select-Databricks -Connection $conn -Table [CData].[Sample].Customers -Where "Country = 'US'" | Remove-Databricks

Modifying Data

The cmdlets make data transformation easy as well as data cleansing. The following example loads data from a CSV file into Databricks, checking first whether a record already exists and needs to be updated instead of inserted. 

Import-Csv -Path C:\My[CData].[Sample].CustomersUpdates.csv | %{
  $record = Select-Databricks -Connection $conn -Table [CData].[Sample].Customers -Where ("_id = `'"+$_._id+"`'")
  if($record){
    Update-Databricks -Connection $conn -Table [CData].[Sample].Customers -Columns @("City","CompanyName") -Values @($_.City, $_.CompanyName) -Where "_id  = `'$_._id`'"
  }else{
    Add-Databricks -Connection $conn -Table [CData].[Sample].Customers -Columns @("City","CompanyName") -Values @($_.City, $_.CompanyName)
  }
}

Copyright (c) 2024 CData Software, Inc. - All rights reserved.
Build 24.0.9060