DATABRICKS SQL EDITOR

Query your Databricks warehouses the SSMS way — browse, write, run, all inside VS Code.

An EpicData tool · Turning Data into Insights

Unofficial extension. Not affiliated with, endorsed by, or sponsored by Databricks, Inc. "Databricks" is a trademark of Databricks, Inc., used here only to describe compatibility.

WHAT IT DOES

Databricks SQL Editor turns VS Code into a lightweight SQL client for Databricks SQL Warehouses. Browse your Unity Catalog objects in a tree, write queries with real metadata-aware autocomplete, press F5, and see results in a grid — without leaving the editor or opening the Databricks web UI.

It talks to the warehouse over the official Statement Execution REST API and signs in with Microsoft Entra ID or a personal access token. No drivers, no extra runtime.

FEATURES

• Object Explorer — browse Catalogs → Schemas → Tables → Columns from the Databricks icon in the activity bar.
• Query editor — write SQL and press F5 or Ctrl+Enter to run. Runs your selection, or just the statement under the cursor when nothing is selected.
• Results grid — results appear in the bottom panel with a Results / Messages split, row count and elapsed time.
• Metadata autocomplete — catalog, schema, table and column suggestions read from your warehouse. Type a . to drill in: catalog. → schemas, catalog.schema. → tables, alias. → columns. Keeps working even while a query is half-typed or has errors, and never interrupts you with a sign-in prompt.
• Databricks-aware syntax — a dedicated Databricks SQL language mode understands backtick-quoted identifiers (`my catalog`.`schema`.`table`), so you don't get false "syntax error" squiggles from the SQL Server tooling. Keywords, strings and functions are colour-coded.
• Right-click actions — Select Top 1000 on a table, New Query on a connection.
• Current catalog / schema — set a default catalog (and schema) from the status bar; it's applied to every query so unqualified names (SELECT * FROM my_table) resolve, autocomplete uses it, and a USE CATALOG / USE SCHEMA in your query updates it automatically.
• Multiple connections — switch the active connection from the status bar; each new query opens in its own dbquery-01, dbquery-02, … tab.
• Secure sign-in — Microsoft Entra ID (via VS Code's built-in Microsoft account) or a personal access token stored in VS Code's secret storage.

HOW TO USE

1. Open the Databricks SQL view (the Databricks icon in the activity bar) and click ＋ Add Connection.
2. Enter your Host (e.g. adb-1234567890.12.azuredatabricks.net), your Warehouse ID or HTTP Path, and choose Microsoft Entra ID (recommended) or a Personal Access Token. For Entra, optionally add your tenant ID.
3. Expand the connection to browse catalogs, schemas, tables and columns.
4. Open a New Query (or right-click a table → Select Top 1000), write your SQL, and press F5 / Ctrl+Enter.
5. Select part of a script to run only that part; with nothing selected, only the statement your cursor is in runs (statements are separated by ;).
6. Switch the active connection any time from the status bar at the bottom, and set the current catalog / schema from the status-bar item next to it (so you can query unqualified table names).

Setting: dbsql.rowLimit (default 10000) caps the rows fetched per query; larger result sets are marked (truncated).

INSTALLATION

Install the packaged .vsix:

• UI: Extensions panel → … menu → Install from VSIX… → select the file.
• CLI: code --install-extension databricks-sql-editor-<version>.vsix

Then reload VS Code and open the Databricks SQL view to add a connection.

CHANGELOG

1.5.0

• List Columns — right-click a table or view → List Columns in One Line or List Columns in a List to insert its backtick-quoted column names at the cursor in the active query.

1.4.4

• Fix "Invalid IP address: undefined" — the direct-DNS lookup now returns the correct shape when Node uses Happy Eyeballs (autoSelectFamily, which calls the lookup with all: true).

1.4.3

• Fix getaddrinfo ENOTFOUND on some Windows setups — the extension now resolves the workspace host via direct DNS (A/AAAA) and connects to that address, bypassing an OS resolver path that fails for azuredatabricks.net hosts even though direct DNS works.

1.4.2

• The workspace host is now sanitized (protocol, path and any stray/invisible characters stripped) before every request.

1.4.1

• View Definition for functions — right-click a function → View Definition to show its definition (via DESCRIBE FUNCTION EXTENDED).

1.4.0

• Tables / Views / Functions folders — each schema now groups its objects into Tables, Views and Functions subfolders in the Object Explorer. Views support Select Top 1000 and View Definition just like tables.

1.3.0

• View Definition — right-click a table or view in the Object Explorer → View Definition to open its CREATE … statement (via SHOW CREATE TABLE) in a new SQL tab.

1.2.1

• Export to CSV — an Export CSV button in the results panel saves the full result of the last query to a flat CSV file, ignoring the dbsql.rowLimit (normal query runs still respect the limit). Streams all rows via the warehouse's external-links path, so large exports work. Every field is double-quoted and separated by ; (semicolon).

1.1.2

• New query files use the .sql extension (opened in Databricks SQL mode, so no false T-SQL squiggles).

1.1.1

• New queries save to a real, user-accessible file. New Query / Select Top 1000 now create a dbquery-NN.sql file in your workspace folder (or a "Databricks Queries" folder in your home directory), so Ctrl+S saves it there instead of prompting for the C:\ root.

1.1.0

• Current catalog / schema — a status-bar item shows the active catalog (and schema); click it to change. The current context is applied to every query (so unqualified names like SELECT * FROM my_table resolve), autocomplete uses it for unqualified tables and columns, and running USE CATALOG / USE SCHEMA updates it automatically.

1.0.0

• First stable release. SSMS-style client for Databricks SQL Warehouses: Object Explorer (Catalogs → Schemas → Tables → Columns), query editor running the selection or the statement under the cursor (F5 / Ctrl+Enter), results grid with row count and timing, metadata-aware autocomplete that keeps working on incomplete queries, Databricks-aware syntax highlighting and colours, named dbquery-NN tabs, and Microsoft Entra ID / personal-access-token sign-in. Zero runtime dependencies.

0.4.x

• Full-colour Marketplace icon and a Databricks-style activity-bar icon with a database badge.
• Named query tabs — new queries open as dbquery-01, dbquery-02, …
• Run selection or current statement — F5 / Ctrl+Enter now run only the selected text, or the single statement under the cursor, instead of the whole file.

0.3.x

• Databricks-aware highlighting — a self-authored grammar gives keywords, strings, functions and identifiers their own colours (functions stay yellow, including keyword-style functions like CAST( and LEFT().

0.2.x

• Metadata autocomplete — catalogs, schemas, tables and columns from the warehouse, context-aware by clause and dotted qualifier, resilient to broken/half-typed SQL, with no sign-in prompts while typing and no noisy word suggestions.

0.1.0

• First release. Connection manager (Microsoft Entra ID / personal access token), Object Explorer, SQL query editor with F5 execution, results grid in the bottom panel, and Select Top 1000.

ABOUT EPICDATA

EpicData is a Belgian data consultancy helping organisations build a data-driven culture — from strategy and governance to platforms, engineering and visualisation.

epicdata.be · Data doet dat.

LICENSE

© 2026 EpicData. Free to download and use (personal or commercial) under the EpicData End-User License Agreement. EpicData retains ownership; modifying, reverse-engineering, reselling, or redistributing the Software outside the official channels is not permitted. See LICENSE.txt.