NebulaDB

NebulaDB is a high-performance, in-memory database engine written in modern C++. It supports a modular SQL-like query system, B+ Tree indexing, transaction management, concurrency control, and both command-line and REST API interfaces.

Features

• SQL-like shell interface (CREATE TABLE, INSERT INTO, SELECT)
• In-memory row and table storage
• B+ Tree index structure
• ACID-style transaction scaffolding with lock manager
• Basic concurrency with thread pool support
• REST API server (optional via Crow)
• Modular file structure and clean header separation
• Unit tests and benchmarking support

Directory Structure

src/          - All C++ source files (core, engine, storage, parser, api)
include/      - All header files for each subsystem
test/         - Unit tests for core components
scripts/      - SQL seed scripts and benchmarking tools
docs/         - Developer notes and roadmap
CMakeLists.txt - Build configuration (CMake)
Makefile      - Optional alternative build script
README.md     - Project overview and instructions

Prerequisites

• C++17-compatible compiler (GCC or Clang)
• CMake (or GNU Make)
• Crow C++ library (if building REST API)

Building and Running

To build and run the command-line shell:

cmake .
make
./nebuladb

Alternatively:

make
./nebuladb

To build and run the unit tests:

make test

Example Shell Commands

CREATE TABLE users id name
INSERT INTO users 1 Alice
INSERT INTO users 2 Bob
SELECT users

REST API (Optional)

To run the REST server:

• Link against Crow in your CMakeLists or Makefile
• In main.cpp, run:

RESTServer::start(db, 8080);

• Send POST requests with raw SQL to http://localhost:8080/query

Roadmap

• Core table/field architecture ✔
• B+ Tree indexing ✔
• SQL parser and shell ✔
• REST API interface ✔
• Transaction manager and lock control ✔
• WHERE clause support
• Joins and aggregation
• Persistent file-backed storage
• Query planner/optimizer engine

License

This project is licensed under the MIT License.