Use the Azure Cosmos DB ODBC driver to connect to BI and data analytics tools


This article walks you through installing and using the Azure Cosmos DB ODBC driver to create normalized tables and views for your Azure Cosmos DB data. You can query the normalized data with SQL queries, or import the data into Power BI or other BI and analytics software to create reports and visualizations.

Azure Cosmos DB is a schemaless database, which enables rapid application development and lets you iterate on data models without being confined to a strict schema. A single Azure Cosmos DB database can contain JSON documents of various structures. To analyze or report on this data, you might need to flatten the data to fit into a schema.

The ODBC driver normalizes Azure Cosmos DB data into tables and views that fit your data analytics and reporting needs. The normalized schemas let you use ODBC-compliant tools to access the data. The schemas have no effect on the underlying data, and don't require developers to adhere to them. The ODBC driver helps make Azure Cosmos DB databases useful for data analysts and development teams.

You can do SQL operations against the normalized tables and views, including group by queries, inserts, updates, and deletes. The driver is ODBC 3.8 compliant and supports ANSI SQL-92 syntax.


Consider using Azure Synapse Link for Azure Cosmos DB to create tables and views for your data. Synapse Link has distinct performance benefits for large datasets over the ODBC driver. You can also connect the normalized Azure Cosmos DB data to other software solutions, such as SQL Server Integration Services (SSIS), QlikSense, Tableau and other analytics software, BI, and data integration tools. You can use those solutions to analyze, move, transform, and create visualizations with your Azure Cosmos DB data.


  • Connecting to Azure Cosmos DB with the ODBC driver is currently supported for Azure Cosmos DB for NoSQL only.
  • The current ODBC driver doesn't support aggregate pushdowns, and has known issues with some analytics tools. Until a new version is released, you can use one of the following alternatives:

Install the ODBC driver and connect to your database

  1. Download the drivers for your environment:

    Installer Supported operating systems
    Microsoft Azure Cosmos DB ODBC 64-bit.msi for 64-bit Windows 64-bit versions of Windows 8.1 or later, Windows 8, Windows 7. 64-bit versions of Windows Server 2012 R2, Windows Server 2012, and Windows Server 2008 R2. Windows 10. Windows Server 2012 R2. Windows Server 2016
    Microsoft Azure Cosmos DB ODBC 32x64-bit.msi for 32-bit on 64-bit Windows 64-bit versions of Windows 8.1 or later, Windows 8, Windows 7, Windows XP, Windows Vista. 64-bit versions of Windows Server 2012 R2, Windows Server 2012, Windows Server 2008 R2, and Windows Server 2003.
    Microsoft Azure Cosmos DB ODBC 32-bit.msi for 32-bit Windows 32-bit versions of Windows 8.1 or later, Windows 8, Windows 7, Windows XP, and Windows Vista.
  2. Run the .msi file locally, which starts the Microsoft Azure Cosmos DB ODBC Driver Installation Wizard.

  3. Complete the installation wizard using the default input.

  4. After the driver installs, type ODBC Data sources in the Windows search box, and open the ODBC Data Source Administrator.

  5. Make sure that the Microsoft Azure DocumentDB ODBC Driver is listed on the Drivers tab.

    Screenshot of the ODBC Data Source Administrator window.

  6. Select the User DSN tab, and then select Add to create a new data source name (DSN). You can also create a System DSN.

  7. In the Create New Data Source window, select Microsoft Azure DocumentDB ODBC Driver, and then select Finish.

  8. In the DocumentDB ODBC Driver DSN Setup window, fill in the following information:

    Screenshot of the domain name server (DNS) setup window.

    • Data Source Name: A friendly name for the ODBC DSN. This name is unique to this Azure Cosmos DB account.
    • Description: A brief description of the data source.
    • Host: The URI for your Azure Cosmos DB account. You can get this information from the Keys page in your Azure Cosmos DB account in the Azure portal.
    • Access Key: The primary or secondary, read-write or read-only key from the Azure Cosmos DB Keys page in the Azure portal. It's best to use the read-only keys, if you use the DSN for read-only data processing and reporting.

    To avoid an authentication error, use the copy buttons to copy the URI and key from the Azure portal.

    Screenshot of the Azure Cosmos DB Keys page.

    • Encrypt Access Key for: Select the best choice, based on who uses the machine.
  9. Select Test to make sure you can connect to your Azure Cosmos DB account.

  10. Select Advanced Options and set the following values:

    • REST API Version: Select the REST API version for your operations. The default is 2015-12-16.

      If you have containers with large partition keys that need REST API version 2018-12-31, type 2018-12-31, and then follow the steps at the end of this procedure.

    • Query Consistency: Select the consistency level for your operations. The default is Session.

    • Number of Retries: Enter the number of times to retry an operation if the initial request doesn't complete due to service rate limiting.

    • Schema File: If you don't select a schema file, the driver scans the first page of data for each container to determine its schema, called container mapping, for each session. This process can cause long startup time for applications that use the DSN. It's best to associate a schema file to the DSN.

      • If you already have a schema file, select Browse, navigate to the file, select Save, and then select OK.

      • If you don't have a schema file yet, select OK, and then follow the steps in the next section to create a schema definition. After you create the schema, come back to this Advanced Options window to add the schema file.

After you select OK to complete and close the DocumentDB ODBC Driver DSN Setup window, the new User DSN appears on the User DSN tab of the ODBC Data Source Administrator window.

Screenshot that shows the new User D S N on the User D S N tab.

Edit the Windows registry to support REST API version 2018-12-31

If you have containers with large partition keys that need REST API version 2018-12-31, follow these steps to update the Windows registry to support this version.

  1. In the Windows Start menu, type regedit to find and open the Registry Editor.

  2. In the Registry Editor, navigate to the path Computer\HKEY_LOCAL_MACHINE\SOFTWARE\ODBC\ODBC.INI.

  3. Create a new subkey with the same name as your DSN, such as Contoso Account ODBC DSN.

  4. Navigate to the new Contoso Account ODBC DSN subkey, and right-click to add a new String value:

    • Value Name: IgnoreSessionToken

    • Value data: 1

    Screenshot that shows the Windows Registry Editor settings.

Create a schema definition

There are two types of sampling methods you can use to create a schema: container mapping or table-delimiter mapping. A sampling session can use both sampling methods, but each container can use only one of the sampling methods. Which method to use depends on your data's characteristics.

  • Container mapping retrieves the data on a container page to determine the data structure, and transposes the container to a table on the ODBC side. This sampling method is efficient and fast when the data in a container is homogenous.

  • Table-delimiter mapping provides more robust sampling for heterogeneous data. This method scopes the sampling to a set of attributes and corresponding values.

    For example, if a document contains a Type property, you can scope the sampling to the values of this property. The end result of the sampling is a set of tables for each of the Type values you specified. Type = Car produces a Car table, while Type = Plane produces a Plane table.

To define a schema, follow these steps. For the table-delimiter mapping method, you take extra steps to define attributes and values for the schema.

  1. On the User DSN tab of the ODBC Data Source Administrator window, select your Azure Cosmos DB User DSN Name, and then select Configure.

  2. In the DocumentDB ODBC Driver DSN Setup window, select Schema Editor.

    Screenshot that shows the Schema Editor button in the D S N Setup window.

  3. In the Schema Editor window, select Create New.

  4. The Generate Schema window displays all the collections in the Azure Cosmos DB account. Select the checkboxes next to the containers you want to sample.

  5. To use the container mapping method, select Sample.

    Or, to use table-delimiter mapping, take the following steps to define attributes and values for scoping the sample.

    1. Select Edit in the Mapping Definition column for your DSN.

    2. In the Mapping Definition window, under Mapping Method, select Table Delimiters.

    3. In the Attributes box, type the name of a delimiter property in your document that you want to scope the sampling to, for instance, City. Press Enter.

    4. If you want to scope the sampling to certain values for the attribute you entered, select the attribute, and then enter a value in the Value box, such as Seattle, and press Enter. You can add multiple values for attributes. Just make sure that the correct attribute is selected when you enter values.

    5. When you're done entering attributes and values, select OK.

    6. In the Generate Schema window, select Sample.

  6. In the Design View tab, refine your schema. The Design View represents the database, schema, and table. The table view displays the set of properties associated with the column names, such as SQL Name and Source Name.

    For each column, you can modify the SQL name, the SQL type, SQL length, Scale, Precision, and Nullable as applicable.

    You can set Hide Column to true if you want to exclude that column from query results. Columns marked Hide Column = true aren't returned for selection and projection, although they're still part of the schema. For example, you can hide all of the Azure Cosmos DB system required properties that start with _. The id column is the only field you can't hide, because it's the primary key in the normalized schema.

  7. Once you finish defining the schema, select File > Save, navigate to the directory to save in, and select Save.

  8. To use this schema with a DSN, in the DocumentDB ODBC Driver DSN Setup window, select Advanced Options. Select the Schema File box, navigate to the saved schema, select OK and then select OK again. Saving the schema file modifies the DSN connection to scope to the schema-defined data and structure.

Create views

Optionally, you can define and create views in the Schema Editor as part of the sampling process. These views are equivalent to SQL views. The views are read-only, and scope to the selections and projections of the defined Azure Cosmos DB SQL query.

Follow these steps to create a view for your data:

  1. On the Sample View tab of the Schema Editor window, select the containers you want to sample, and then select Add in the View Definition column.

    Screenshot of creating a view within the driver.

  2. In the View Definitions window, select New. Enter a name for the view, for example EmployeesfromSeattleView, and then select OK.

  3. In the Edit view window, enter an Azure Cosmos DB query, for example:

    SELECT c.City, c.EmployeeName, c.Level, c.Age, c.Manager FROM c WHERE c.City = "Seattle"

  4. Select OK.

    Screenshot of adding a query when creating a view.

You can create as many views as you like. Once you're done defining the views, select Sample to sample the data.


The query text in the view definition should not contain line breaks. Otherwise, we will get a generic error when previewing the view.

Query with SQL Server Management Studio

Once you set up an Azure Cosmos DB ODBC Driver User DSN, you can query Azure Cosmos DB from SQL Server Management Studio (SSMS) by setting up a linked server connection.

  1. Install SQL Server Management Studio and connect to the server.

  2. In the SSMS query editor, create a linked server object for the data source by running the following commands. Replace DEMOCOSMOS with the name for your linked server, and SDS Name with your data source name.

    USE [master]
    EXEC master.dbo.sp_addlinkedserver @server = N'DEMOCOSMOS', @srvproduct=N'', @provider=N'MSDASQL', @datasrc=N'SDS Name'
    EXEC master.dbo.sp_addlinkedsrvlogin @rmtsrvname=N'DEMOCOSMOS', @useself=N'False', @locallogin=NULL, @rmtuser=NULL, @rmtpassword=NULL

To see the new linked server name, refresh the linked servers list.

Screenshot showing a linked server in S S M S.

To query the linked database, enter an SSMS query. In this example, the query selects from the table in the container named customers:


Execute the query. The results should look similar to the following output:

attachments/  1507476156    521 Bassett Avenue, Wikieup, Missouri, 5422   "2602bc56-0000-0000-0000-59da42bc0000"   2015-02-06T05:32:32 +05:00 f1ca3044f17149f3bc61f7b9c78a26df
attachments/  1507476156    167 Nassau Street, Tuskahoma, Illinois, 5998   "2602bd56-0000-0000-0000-59da42bc0000"   2015-06-16T08:54:17 +04:00 f75f949ea8de466a9ef2bdb7ce065ac8
attachments/  1507476156    885 Strong Place, Cassel, Montana, 2069       "2602be56-0000-0000-0000-59da42bc0000"   2015-03-20T07:21:47 +04:00 ef0365fb40c04bb6a3ffc4bc77c905fd
attachments/  1507476156    515 Barwell Terrace, Defiance, Tennessee, 6439     "2602c056-0000-0000-0000-59da42bc0000"   2014-10-16T06:49:04 +04:00      e913fe543490432f871bc42019663518
attachments/  1507476156    570 Ruby Street, Spokane, Idaho, 9025       "2602c156-0000-0000-0000-59da42bc0000"   2014-10-30T05:49:33 +04:00 e53072057d314bc9b36c89a8350048f3

View your data in Power BI Desktop

You can use your DSN to connect to Azure Cosmos DB with any ODBC-compliant tools. This procedure shows you how to connect to Power BI Desktop to create a Power BI visualization.

  1. In Power BI Desktop, select Get Data.

    Screenshot showing Get Data in Power B I Desktop.

  2. In the Get Data window, select Other > ODBC, and then select Connect.

    Screenshot that shows choosing ODBC data source in Power B I Get Data.

  3. In the From ODBC window, select the DSN you created, and then select OK.

    Screenshot that shows choosing the D S N in Power B I Get Data.

  4. In the Access a data source using an ODBC driver window, select Default or Custom and then select Connect.

  5. In the Navigator window, in the left pane, expand the database and schema, and select the table. The results pane includes the data that uses the schema you created.

    Screenshot of selecting the table in Power B I Get Data.

  6. To visualize the data in Power BI desktop, select the checkbox next to the table name, and then select Load.

  7. In Power BI Desktop, select the Data tab on the left of the screen to confirm your data was imported.

  8. Select the Report tab on the left of the screen, select New visual from the ribbon, and then customize the visual.


  • Problem: You get the following error when trying to connect:

    [HY000]: [Microsoft][Azure Cosmos DB] (401) HTTP 401 Authentication Error: {"code":"Unauthorized","message":"The input authorization token can't serve the request. Please check that the expected payload is built as per the protocol, and check the key being used. Server used the following payload to sign: 'get\ndbs\n\nfri, 20 jan 2017 03:43:55 gmt\n\n'\r\nActivityId: 9acb3c0d-cb31-4b78-ac0a-413c8d33e373"}

    Solution: Make sure the Host and Access Key values you copied from the Azure portal are correct, and retry.

  • Problem: You get the following error in SSMS when trying to create a linked Azure Cosmos DB server:

    Msg 7312, Level 16, State 1, Line 44
    Invalid use of schema or catalog for OLE DB provider "MSDASQL" for linked server "DEMOCOSMOS". A four-part name was supplied, but the provider does not expose the necessary interfaces to use a catalog or schema.

    Solution: A linked Azure Cosmos DB server doesn't support four-part naming.

Next steps