Databricks Partner Connect targets Power BI, not Excel. The ODBC path requires the HTTP path, not just a hostname. Every option explained.
Databricks now has an official Excel Add-in, in public preview since April 2026. It lets you browse Unity Catalog tables, write SQL, and import results directly into Excel without configuring an ODBC driver. The older path, Partner Connect, does not connect to Excel; it launches Power BI Desktop. For teams not yet on the official add-in, the Simba ODBC driver works against a SQL warehouse but requires per-machine setup and the HTTP path that most guides forget to mention. A CSV download from the SQL editor covers one-off pulls with no setup at all. ODBC works on Windows and Mac, but Mac requires manual config file editing and does not reach Excel on the web. Here is every option.
Method 1: Databricks Excel Add-in (official, public preview)
Databricks released an official Excel Add-in in April 2026, currently in public preview. Install it from Microsoft AppSource or from the setup page in your Databricks workspace under Integrations, Excel Add-in. Once installed, the add-in appears in the Excel ribbon. Sign in with your Databricks workspace URL and your credentials. You can then browse Unity Catalog tables by catalog, schema, and table, write a SQL query, preview results, and import them into the active sheet. The add-in also supports writing data back to Unity Catalog tables, which the ODBC path does not.
Full setup steps are in the Databricks documentation under Integrations, Connect to Databricks from Microsoft Excel. The Microsoft Learn version covers Azure Databricks specifically; the AWS and GCP docs pages cover the same add-in for those platforms.
The catch: public preview means features and behaviour can change. Unity Catalog is required; the add-in does not support Hive Metastore-only workspaces. It targets Windows and Mac desktop Excel; web Excel support is not confirmed in the preview documentation. SQL warehouse must be running to accept connections.
Method 2: Simba ODBC driver and Get Data from ODBC
Databricks publishes the Simba Databricks ODBC driver through their drivers documentation page. Download the version for your operating system, install it, then create a DSN in your ODBC administrator.
The DSN requires two fields that are specific to Databricks and differ from most ODBC setups: the Server Hostname (your workspace URL, typically in the format adb-123456789.azuredatabricks.net or the equivalent for AWS/GCP) and the HTTP Path, which identifies the specific SQL warehouse. Find the HTTP path in your Databricks workspace under SQL Warehouses, select the warehouse, then the Connection Details tab. It looks like /sql/1.0/warehouses/abc123. For authentication, use a personal access token: enter token as the username and your generated token as the password.
On Windows: once the DSN is created in the 64-bit ODBC Data Source Administrator, go to Excel, then Data, then Get Data, then From Other Sources, then From ODBC, select the DSN, and navigate to the catalog, schema, and table you want. Power Query loads the result.
On Mac: install iODBC via Homebrew or a downloaded package, then manually edit ~/.odbc.ini to define the DSN (hostname, HTTP path, token) and ~/.odbcinst.ini to register the Simba driver. Excel for Mac should then pick up the DSN. Test with the iODBC administrator tool before trying Excel. Mac setup is significantly more involved than Windows.
The catch: every machine that needs the connection requires the driver and DSN setup. ODBC does not work in Excel on the web. SQL warehouses must be running to accept connections; a stopped warehouse causes a connection error until it warms up.
A note on Partner Connect: it does not target Excel
Databricks Partner Connect is a feature in the workspace that pre-configures connections to partner tools including Power BI, Tableau, and others. If you have seen documentation suggesting Partner Connect can help with Excel, note that the BI partner it launches is Power BI Desktop, not Excel. Partner Connect does not produce an Excel connection. If you need both Power BI and Excel connections to Databricks, they are configured separately: Power BI via Partner Connect or its native Databricks connector, Excel via ODBC or an add-in.
Method 3: SQL editor or notebook results CSV download
For a one-off pull without any driver setup: open the Databricks SQL editor, write your query, run it, and download the results as CSV using the download option in the results panel. Notebook outputs can also be displayed as tables with a download option depending on the display mode. Open the CSV in Excel.
The catch: this is a frozen snapshot. The SQL editor caps displayed results (typically 10,000 rows); larger results are truncated in the display even if the query ran on the full dataset. For complete large result sets, write the results to a Delta table or a file in your storage account and download from there.
Method 4: Brooked for Excel (scheduled, Mac and web Excel)
Brooked for Excel connects to Databricks SQL warehouses over an authenticated API, so there is no ODBC driver to install on each machine. It runs inside Excel's ribbon and works in desktop Excel on Windows, desktop Excel on Mac, and Excel on the web, covering the cases that ODBC cannot reach.
Honest framing: Excel support is in beta. Manual imports and the AI analyst work today; scheduled refresh is rolling out through the beta. The Google Sheets connector for Databricks is fully live if your team uses Sheets.
Setting up the Databricks connection in Brooked
Open Brooked in the Excel ribbon, select Databricks as the source, and enter your workspace URL and a personal access token. Select the SQL warehouse to use, then browse the catalog and schema to pick a table, or enter a custom SQL query. Import the results into the active worksheet.
Brooked AI, DatabricksMethod comparison
| Method | Cost | Works on | Refresh | Best for |
|---|---|---|---|---|
| Databricks Excel Add-in (official) | Free (public preview) | Windows and Mac desktop Excel | Manual refresh in add-in | Unity Catalog users wanting the official Databricks tool |
| Databricks ODBC driver + Get Data | Free (driver download required) | Windows Excel, Mac Excel (manual config) | Manual refresh in Excel | SQL warehouse access from a managed Windows environment |
| SQL editor or notebook CSV download | Free | Any Excel | Manual re-download | One-off analyst queries or notebook output |
| Brooked for Excel | Free during beta | Desktop, Mac & web Excel | Scheduled (rolling out in beta) | Live Databricks data without driver setup |
What data teams build with this
- Product analytics reporting: DAU, retention, and funnel metrics from the gold-layer tables refreshed into an Excel dashboard
- Experiment results: cohort analysis output from a Databricks notebook pulled into Excel for statistical review
- Finance reporting: aggregated revenue and cost tables from the lakehouse landing alongside budget models in Excel
- Data quality monitoring: row counts, null rates, and schema drift metrics pulled on a schedule
- Marketing performance: campaign attribution data from Databricks joined to spend data already in the workbook
Troubleshooting common issues
Related Articles
- How to Connect Snowflake to Excel
- How to Connect Amazon Redshift to Excel
- How to Connect Databricks to Google Sheets
Frequently asked questions
Can I use Partner Connect to connect Databricks to Excel?
No. Databricks Partner Connect is designed to connect Databricks to partner BI tools and data tools. The BI partners listed include Power BI, Tableau, and others, but not Excel. Partner Connect launches Power BI Desktop with the connection pre-configured. It does not create an Excel connection. For Excel, you need the ODBC driver or a third-party add-in.
What is the HTTP path and where do I find it?
The HTTP path is a unique identifier for your Databricks SQL warehouse, separate from the server hostname. Without it, the ODBC connection cannot route to the right compute resource. Find it in the Databricks workspace under SQL Warehouses, select your warehouse, then the Connection Details tab. It looks like /sql/1.0/warehouses/abc123def456. You need both the Server Hostname and the HTTP path to configure the DSN.
How do I connect Databricks to Excel on Mac?
Download the Simba Databricks ODBC driver for Mac from the Databricks drivers page. Install it, then use iODBC or unixODBC as the driver manager. Edit the odbc.ini and odbcinst.ini config files manually to register the DSN with your server hostname, HTTP path, and a personal access token. Excel for Mac can then pick up the DSN. The Mac path is significantly more involved than Windows and requires command-line work.
How do I authenticate the ODBC connection to Databricks?
The most common authentication method for ODBC is a personal access token. Generate one in the Databricks workspace under Settings, then Developer, then Access tokens. Enter 'token' as the username and the token string as the password in the DSN configuration. OAuth-based authentication is also supported in newer driver versions but requires additional configuration.
Does the Databricks ODBC connection work in Excel on the web?
No. ODBC requires a locally installed driver and a DSN configured on the machine running Excel. Excel on the web runs in the browser and cannot reach a local ODBC DSN. To get Databricks data into web Excel, you need an add-in that authenticates over the internet, such as Brooked.
What is a SQL warehouse in Databricks, and do I need one for Excel?
A SQL warehouse (previously called SQL endpoint) is the compute cluster that processes SQL queries in Databricks. Unlike all-purpose clusters, SQL warehouses are optimised for BI and query workloads and are the recommended compute type for connecting Excel or BI tools. You can also connect to an all-purpose cluster using its HTTP path, but SQL warehouses are designed for this use case and tend to be more cost-effective for repeated query workloads.
Does Databricks have an official Excel add-in?
Yes, as of April 2026 Databricks has an official Excel Add-in in public preview. It connects to your Databricks workspace, lets you browse Unity Catalog tables, write SQL queries, and import results into Excel without setting up an ODBC driver. It also supports writing data back to Unity Catalog tables. Install it from Microsoft AppSource or from the Integrations section of your Databricks workspace. Unity Catalog is required; Hive Metastore-only workspaces are not supported in the preview.
Connect Databricks to Excel without a driver install.
Brooked for Excel is in beta and free to try. Works on Mac and web Excel where ODBC cannot. Using Google Sheets instead? Read the Sheets version of this guide.
See Brooked for Excel →
