Java SQL Engine

A foundational project to build a lightweight, relational database engine from scratch in Java. The primary goal is to gain a deep, systems-level understanding of database internals, including storage, indexing, concurrency, and recovery.

This repository also contains a detailed developer log that chronicles the design decisions, challenges, and learning process throughout the project's development.

🏛️ Architecture Overview

The engine is designed with a classic, layered database architecture.

Client -> Parser -> Planner -> Executor -> Transaction & Storage Manager

• SQL Parser: Converts SQL strings into an Abstract Syntax Tree (using ANTLR).
• Query Planner: Translates the AST into a logical plan of operators.
• Execution Engine: Executes the plan (e.g., SeqScan, IndexScan).
• Transaction Manager: Ensures ACID properties using locking and logging.
• Storage Manager: Manages data on disk through a buffer pool.
  • Buffer Manager: Caches disk pages in memory using an LRU policy.
  • Disk Manager: Handles the physical I/O of reading/writing pages to table files.
• Index Manager: Manages B+-Tree indexes for fast lookups.

✅ Project Roadmap

• Phase 1: The Storage Layer
  • Initial Project Setup
  • Page Class and On-Disk Representation
  • DiskManager for Raw Page I/O
  • BufferPoolManager with LRU Caching
  • Heap File and Tuple Storage
• Phase 2: Basic Query Execution
  • SQL Parser (ANTLR)
  • SeqScan (Sequential Scan) Executor
• Phase 3: Indexing
  • B+-Tree Implementation
  • IndexScan Executor
• Phase 4: Concurrency & Recovery
  • Transaction Manager
  • Write-Ahead Logging (WAL)
  • Strict Two-Phase Locking (2PL)

🚀 Getting Started

This project is built using Java 21 and Maven.

# Clone the repository
git clone https://github.com/lokiT26/java-sql-engine.git

# Navigate into the project directory
cd java-sql-engine

# Build the project and run tests
mvn clean install