Skip to main content

Documentation Index

Fetch the complete documentation index at: https://datost.com/docs/llms.txt

Use this file to discover all available pages before exploring further.

Datost connects to Databricks through the SQL Statement Execution API. Once the warehouse is connected, Datost can list tables in INFORMATION_SCHEMA, read column types, and run read-only queries on behalf of Slack users.

​
What it unlocks

  • Natural-language querying over any catalog and schema your warehouse can reach, including Unity Catalog.
  • Schema introspection across tables and views (the VIEW table type is recognized), powered by a single INFORMATION_SCHEMA pass so large workspaces load without N+1 fan-out.
  • Scoped access: pin the connection to a specific catalog and default schema, or leave them blank for workspace defaults.

​
Prerequisites

You need admin access to the Databricks workspace to create a SQL Warehouse and mint a token.
  • Workspace host in the form dbc-xxxxxxxx-xxxx.cloud.databricks.com (no https:// prefix).
  • SQL Warehouse HTTP path in the form /sql/1.0/warehouses/{id}. Datost validates this pattern and extracts the warehouse ID.
  • Personal Access Token (PAT) from User Settings → Developer → Access Tokens, or a service-principal token with equivalent grants.
  • Optional: a default catalog (for example main or samples) to scope the connection.

​
Connect in the admin panel

1

Open Add Data Source

In the Datost admin, go to Data Sources → Add Data Source and pick Databricks.
2

Enter the workspace host

Paste the workspace URL, for example dbc-xxxxxxxx-xxxx.cloud.databricks.com. This is the host Datost will call for the SQL Statement Execution API.
3

Paste the HTTP path

From the SQL Warehouse’s Connection Details tab, copy the HTTP path (for example /sql/1.0/warehouses/abc123...).
4

Add the access token

Paste the PAT (it begins with dapi). Datost stores it encrypted and only uses it to execute statements against this warehouse.
5

Optionally pin a catalog

Set a default catalog if you want to scope the connection. Leave blank to use the workspace default.
6

Test and save

Datost runs SELECT 1 to verify the connection before saving. A successful test confirms the token, host, and warehouse ID are valid.

​
Permissions

The token’s identity needs read access to anything you want Slack users to query:
  • USE CATALOG on each catalog Datost should see.
  • USE SCHEMA on each schema within those catalogs.
  • SELECT on the tables and views you want exposed.
  • Read access to INFORMATION_SCHEMA in each catalog, which Datost uses to list tables and columns.
Datost executes whatever SQL the warehouse authorizes for this token. Scope the principal to read-only grants on only the catalogs and schemas your team should query.

​
Limitations

The integration targets Databricks SQL Warehouses via the /api/2.0/sql/statements endpoint. All-purpose clusters and job clusters are not supported as connection targets.
Both serverless and classic SQL Warehouses work. Serverless starts faster, which makes the first query after an idle period feel more responsive; classic warehouses may cold-start before the 30-second statement wait window.
Statements run with format: JSON_ARRAY and disposition: INLINE. Query results are capped (default 1,000 rows) and flagged as truncated when the cap is hit.
Datost caches up to 50 warehouse client entries for 30 minutes each. Rotating a PAT requires re-saving the data source so the cached entry is invalidated.