GPU-Accelerated Columnar Database in Rust

This project is an attempt to build a GPU-accelerated columnar database from scratch, inspired by systems like Snowflake, using Rust. It leverages object storage (AWS S3-compatible) for data persistence, and GPU processing for fast analytical queries.

Getting Started

Building and Running

1. Clone the repository:

   git clone https://github.com/hamzzy/giga-db
   cd giga-db

2. Build the project:

   cargo build

3. Start the database:

   cargo run

4. Send Queries:

   Use a tool like curl or Postman to send POST requests to http://127.0.0.1:8080/query with the following json: ```json { "query": "SELECT * FROM test WHERE id > 1.0"}

     Example Queries:
   *  `SELECT id, value FROM test WHERE id > 1.0` (to test projection)
   *   `SELECT SUM(id) FROM test` (to test aggregation)
   *   `SELECT * FROM test WHERE id > 1.0` (to test projection with wild card)
   
   

Configuration

This application uses .env to configure environment variables such as credentials, and bucket name. These are parsed during start up.

Contributing

We welcome contributions to this project. To contribute, please:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Implement your changes, write tests, and make sure all the existing tests pass.
4. Submit a pull request with a detailed description of your changes.