DuckPaul

SQL transformations powered by DuckDB. Write plain SQL, test locally, ship to production.

The Problem

Most SQL transformation tools either lock you into a proprietary DSL, require a heavy runtime, or charge you for features that should be free. DuckPaul uses DuckDB under the hood — it's fast, it runs anywhere, and your transforms are just plain SQL files.

How It Will Work

1. Write SQL files — each file is a transform. Dependencies are declared with a comment.
2. Test locally — DuckDB means fast local execution with no external database needed.
3. Ship to production — run the same transforms against Postgres, Parquet files, or remote data.

Example Transform

-- depends_on: raw_data
SELECT
    order_id,
    customer_id,
    CAST(amount AS DECIMAL(10,2)) AS amount,
    created_at::DATE AS order_date
FROM raw_data
WHERE status != 'cancelled'

Project Structure

my-transforms/
├── duckpaul.yaml
├── sources/
│   └── raw_data.sql
├── transforms/
│   └── clean_orders.sql
└── tests/
    └── test_orders.sql

Planned Features

• DuckDB-powered local execution (no database server needed)
• Dependency resolution between transforms
• Built-in data testing framework
• Incremental builds for large datasets
• Multiple data source support (Parquet, CSV, Postgres, S3)
• CI/CD integration

Get Early Access

If you're working with SQL transformations and want a lightweight, open-source option, get in touch. We're looking for early users to help shape the tool.