Getting started
This page takes you from a fresh install to your first query. Install omni-sql first: see Installation.
Query a file
Section titled “Query a file”You do not need a database to start.
-
Choose Import file and pick a CSV, JSON or Parquet file. For CSV, check the preview and the delimiter, then confirm.
-
The file becomes a table in the Local DuckDB workspace, listed in the sidebar.
-
In the Local DuckDB SQL tab, write a query and press Ctrl+Enter (Cmd+Enter on macOS):
SELECT region, count(*) AS orders, sum(total) AS revenueFROM ordersGROUP BY regionORDER BY revenue DESC;
Imported tables persist across restarts. See Local analysis for samples, exports and S3.
Connect to a database
Section titled “Connect to a database”- Open the connection switcher in the header (Ctrl+Alt+O) and choose New connection.
- Pick the database type and fill in host, port, database, user and password. Optionally choose a folder, an environment, SSL settings and which schemas to load.
- Choose Test connection, then save and connect.
The sidebar now lists the connection’s schemas, tables, views and functions.
Double-click a table to open a SELECT, or open a new SQL tab and start typing:
completion knows your tables and columns. Connection options are covered in
Connections, and per-database notes in
Database support.
Run and inspect queries
Section titled “Run and inspect queries”- Ctrl+Enter runs the selection, or the statement under the cursor. Ctrl+Shift+Enter runs every statement in the tab.
- Results appear below the editor. Filter, sort, page and export them, and raise the Row limit in the toolbar when you need more rows.
- Explain SQL shows the execution plan where the database supports it.
- Statements that can remove data (
DROP,TRUNCATE,DELETEwithoutWHERE, …) ask for confirmation first.
More in SQL editor.
Join a query result with a file
Section titled “Join a query result with a file”Choose Analyze locally after running a query to copy its result, in full or as a sample, into DuckDB. A new Local DuckDB tab opens, where you can join it with imported files. The database + CSV tutorial walks through a complete example.
Next steps
Section titled “Next steps”- Connect an AI assistant with the MCP integration.
- Query MongoDB natively or with SQL: MongoDB.
- Learn the keyboard shortcuts.