> AI agents: this is one page from Mammoth Analytics documentation. The index of all pages as Markdown is https://docs.mammoth.io/llms.txt. Append `.md` to any docs URL, or send `Accept: text/markdown`, to get Markdown.

# Databricks connector

> Connect Mammoth to Databricks with a personal access token and a SQL warehouse HTTP Path, then import a table or SQL query as a dataset.

The Databricks connector imports data from your Databricks lakehouse into Mammoth as a dataset. Connect to a SQL warehouse using a personal access token, then browse tables across a catalog and its schemas or write custom SQL queries.

## Before you start

You need:

- A running Databricks SQL warehouse
- A Databricks **personal access token** (PAT) with access to the catalog you want to import from
- The warehouse's **HTTP Path**, from **Databricks → SQL Warehouses → {warehouse} → Connection details**
- Your Databricks workspace **Host URL** and the **Catalog Name** you want to query
- Mammoth's IP address added to your Databricks network configuration—see [Network requirements](https://docs.mammoth.io/learn/connectors/network-requirements/)

> **Tip: Personal access token is the only supported auth method.**
> The Databricks connector authenticates with a personal access token only—there's no OAuth or service-principal option. Generate a token from Databricks → User Settings → Developer → Access tokens, and note its expiration date, since PAT expiry is the most common reason a Databricks connection stops working (see Troubleshooting).

## Step 1: Select the Databricks connector

1. In Mammoth, go to **Data Library** and click **+ New Dataset**
2. Select **Connectors**
3. Search for **Databricks** or find it in the **Databases** category
4. Click the **Databricks** tile

## Step 2: Create a connection

Click **New Connection** to open the credential form.

| Field | Description | Example |
| --- | --- | --- |
| **Host URL** | Your Databricks workspace hostname | `dbc-a1b2c3d4-e5f6.cloud.databricks.com` |
| **Port** | Connection port | `443` (default) |
| **HTTP Path** | The SQL warehouse's connection endpoint | `/sql/1.0/warehouses/a1b2c3d4e5f6g7h8` |
| **Personal Access Token** | Your Databricks PAT |—|
| **Catalog Name** | The Unity Catalog to connect to—fixed for this connection | `main` |
| **Display Name** | Optional label for this connection in Mammoth | `Databricks Production` |

### Find your Databricks HTTP Path

In Databricks, go to **SQL Warehouses**, select the warehouse you want to use, and open **Connection details**. Copy the **HTTP Path** value exactly—this field is specific to Databricks and doesn't map to anything in other database connectors, so it's the field most likely to trip up first-time users.

A callout below the fields reminds you to whitelist Mammoth's IP address in your database settings—see [Network requirements](https://docs.mammoth.io/learn/connectors/network-requirements/) for the exact address to use.

There's no cluster picker in this form: the **HTTP Path** determines which warehouse Mammoth queries, and the **Catalog Name** is fixed per connection—connecting to a different catalog means creating a new connection.

After filling in the fields, click **Connect**. Mammoth validates the credentials and saves the connection; if validation fails, the error message appears at the bottom of the form.

## Step 3: Select a Databricks schema and query your data

Once connected, select the connection under **Existing Connections**. Mammoth lists the schemas in your chosen catalog under **Schema**. Click a schema to browse its tables in the **List of tables** panel.

### Browse Databricks tables

Click any table to load a fully qualified preview query, backtick-quoted in Databricks' three-level catalog → schema → table style: ``SELECT * FROM `catalog`.`schema`.`table` ``.

### Write a custom Databricks query

```sql
SELECT order_id, customer_id, amount, order_date
FROM `main`.`sales`.`orders`
WHERE order_date >= current_date() - INTERVAL 90 DAYS
  AND status = 'completed'
```

> **Tip: No AI query generation for Databricks.**
> Databricks doesn't offer the **Describe the data you need...** AI query box. Write or edit SQL directly in the **Enter SQL query** field.

See [Understanding data preview & query](https://docs.mammoth.io/learn/connectors/understanding-data-preview-query/) for a full guide to the preview screen.

## Step 4: Save the Databricks dataset

1. Enter a **dataset name** in the **Save table as** field
2. Check the **Update Method** (the default is **Replace all data**). To change it, click the edit icon, pick an option, and click **Save changes**:
   - **Replace all data**—Replaces all existing data
   - **Add new data**—Adds new data based on a unique sequential column
   - **Replace with new data**—Replaces existing data with new data based on a unique sequential column
   For **Add new data** and **Replace with new data**, also choose a **Key column**—a numeric or date column with incrementing values, such as an ID or a date.
3. Click **Save as dataset**. When the dataset is saved, click **Done**

## Allow Mammoth in your Databricks network configuration

Add Mammoth's IP address to your Databricks workspace's allowed network configuration. See [Network requirements](https://docs.mammoth.io/learn/connectors/network-requirements/) for the current IP address and configuration steps.

## Troubleshoot Databricks connections

### Personal access token is no longer valid

The error "Authentication Failed, Your personal access token is no longer valid. Please check and update the token in the identity configuration." is the most common error for this connector. Databricks personal access tokens expire on a date set by your workspace administrator. To fix it:

1. In Databricks, go to **User Settings → Developer → Access tokens**
2. Generate a new token (or check the expiration date on your existing one)
3. Update the **Personal Access Token** field on your Mammoth connection with the new value

### Connection fails with a valid token

- Double-check the **HTTP Path**—it must come from a currently running SQL warehouse's **Connection details**, not a cluster or job
- Confirm Mammoth's IP address is whitelisted in your Databricks network configuration

### No tables appear after selecting a schema

Your personal access token's associated user may lack access to that catalog or schema in Unity Catalog. Ask your Databricks workspace administrator to grant access.

---
Source: https://docs.mammoth.io/learn/connectors/databricks.md · Updated: 2026-10-13