Open CSV or Parquet files with a DuckDB-backed Notebook interface. Run multiple SQL queries in separate cells, view results in interactive tables, and export data.
Features
1. Interactive Notebook
Smart Execution: Run cells with Cmd+Enter or Shift+Enter (runs and adds a new cell below).
Cell Management:
Drag & Drop: Reorder cells by dragging the handle.
Add/Remove: Insert new cells between existing ones or delete them.
Cancellation: Stop long-running queries with the Stop button.
2. Rich Data Viewer
Interactive Tables: Scrollable results with sticky headers.
JSON Support: Automatically detects and renders JSON objects/arrays in a collapsible tree view.
NULL Handling: Visually distinguishes NULL values from empty strings.
Clickable Links: Automatically detects and links URLs in result cells.
Column Types: Displays DuckDB column types in usage.
3. Supported Formats
CSV: Automatic dialect detection, with specific handling for headers.
Parquet: High-performance native support.
4. Customization & Security
Settings Modal: Configure behavior via the settings gear icon:
Text Wrap: Toggle text wrapping in grid cells.
JSON Parsing: Force JSON parsing for ambiguous columns.
Preview Limit: Set the default number of rows to fetch.
Security: Explicit permission prompts for accessing external files (e.g., read_csv('/abs/path.csv')).
5. Export
One-Click Export: Download cell results as CSV or Parquet files directly from the cell header.
Copy to Clipboard: Copy data formatted for spreadsheet pasting.
Notes
DuckDB runs in the webview via @duckdb/duckdb-wasm, bundled locally in media/duckdb (no CDN required at runtime).
For large datasets, limit your queries; only the first 100 rows are shown by default when opening a file.
