Visual Join Builder

☕ Buy me a coffee

Visual Join Builder is a VS Code extension for building DataFrame joins visually and generating production-ready code for Pandas, DuckDB, and PySpark.

What it does

• Drag DataFrames from the sidebar onto a visual canvas.
• Connect table columns to create joins.
• Change join types from:
  • edge join controls,
  • interactive Venn mode,
  • selected-edge Venn overlay editing.
• See generated code update in real time.
• Preview results from the active Jupyter kernel.
• Insert generated code directly into a notebook or Python editor.

Key features

• Live kernel integration: detects DataFrames from the active/visible Jupyter notebook kernel.
• Manual mode fallback: UI remains usable even when no kernel is active.
• Join graph UX:
  • auto-arranged table drops,
  • auto-fit canvas behavior,
  • edge and node deletion support,
  • floating control panels.
• Interactive Venn editor:
  • clickable left/intersection/right regions,
  • smart mapping to INNER/LEFT/RIGHT/FULL OUTER joins,
  • guidance for unsupported region combinations.
• Code generation engine:
  • Pandas merge chains with collision handling,
  • DuckDB SQL via duckdb.sql(query).df(),
  • PySpark chained .join() and .select().

Commands

• visual-join-builder.open — Open Visual Join Builder.
• visual-join-builder.openFromCell — Open from notebook/Python context.

CodeLens

The extension adds a ✨ Visual Join CodeLens in notebook/Python files when Pandas usage is detected.

Requirements

• VS Code ^1.85.0
• Python/Jupyter extension (ms-toolsai.jupyter) for live kernel table discovery and preview

Packaging

• Native sponsor metadata is configured in package.json (sponsor.url) so VS Code can show the sponsor heart on the extension page.
• Marketplace packaging uses .vscodeignore to include runtime artifacts and exclude development files.

License

See repository license and metadata files for distribution details.