PyRDB

A toy relational-database engine in Python, built to learn how RDBMS internals work.

Status: Schema-driven file storage is implemented; basic inserts and selects succeed.

Features

• Schema-driven: define tables and columns in a JSON file.
• Binary storage: data is written to and read from binary files.
• Simple API: supports INSERT and full-table SELECT.
  • Foreign-key enforcement: not yet supported.
  • Row filtering: not yet supported.

Roadmap

Next steps

DB internals

• Set up primary keys - one b-tree index file per table
  • addition of an index to the binary tree
  • deletion of an index and taking care how to either mantain or update the memory offsets correspomnding to the oter indexes.
• Enforce foreign-key constraints
• Implement WHERE-clause filtering
• Support join operations

SQL parser

Parse SQL commands into functions

DB server

Build a TCP-socket server with:

• Event listener
• SQL-text parser
• Client authentication

Front end

• Develop a minimal web client interface
• Require authentication before quering