Skip to main content

​
Introduction to Databas

Databas is a toy SQLite-like database system implemented in Rust for educational purposes. It demonstrates core database concepts including page-based storage, SQL parsing, and query execution.
This is a work-in-progress educational project. It is not intended for production use.

​
What is Databas?

Databas implements a simplified relational database system that uses many of the same concepts found in SQLite:
  • Page-based storage with a custom binary format
  • SQL parser that handles SELECT, INSERT, and CREATE TABLE statements
  • Command-line interface for interactive queries
  • B-tree data structures for efficient table storage
  • Page cache with clock replacement algorithm

​
Key features

SQL Parser

Hand-written lexer and parser supporting SELECT, INSERT, and CREATE TABLE statements with expressions, WHERE clauses, ORDER BY, and LIMIT.

Page-based Storage

Custom binary format with 4KB pages, magic number validation, and checksums. Supports both leaf and interior B-tree pages.

Type System

Supports INT, FLOAT, and TEXT column types with PRIMARY KEY and NULLABLE constraints.

Interactive CLI

Command-line interface with REPL mode for executing queries and viewing parsed SQL statements.

​
Architecture overview

Databas is organized as a Rust workspace with three main crates:

​
databas_cli

The command-line interface that provides:
  • Interactive REPL with >>> prompt
  • Single query execution mode
  • SQL statement parsing and display

​
databas_core

The core database engine implementing:
  • DatabaseHeader - Database metadata with magic number databas format1\0
  • DiskManager - File I/O operations for reading and writing pages
  • PageCache - In-memory cache with clock replacement and pin guards
  • TablePage - B-tree page structures (leaf and interior nodes)
  • Page checksums for data integrity

​
databas_sql_parser

The SQL parsing layer featuring:
  • Lexer - Tokenizes SQL statements
  • Parser - Builds abstract syntax trees using Pratt parsing
  • Support for expressions, operators, aggregates, and literals
  • Error handling with position tracking

​
Get started

Quickstart

Build from source and run your first queries

Architecture

Deep dive into the database internals

​
Example queries

Databas supports standard SQL syntax for table creation, data insertion, and queries: 
CREATE TABLE products (id INT PRIMARY KEY, name TEXT, price FLOAT);

INSERT INTO products (id, name, price) VALUES (123, 'Cake', 45.67);

SELECT name, price FROM products WHERE price > 10.00 ORDER BY price DESC;

​
Why Databas?

This project is designed for learning:
  • Understand database internals - See how page-based storage actually works
  • Learn Rust - Practical example of systems programming with Rust
  • Study parsing - Hand-written lexer and parser implementation
  • Explore data structures - B-trees, page caches, and disk management
The codebase is intentionally kept simple and readable to serve as a learning resource.

Build docs developers (and LLMs) love

Get started for freeTalk to us