Pesapal RDBMS

Problem

Understanding how relational databases work under the hood means building the pieces most developers never touch: the parser, planner, and storage engine.

Approach

Built a relational database engine from scratch in Python using only the standard library, for the Pesapal JDEV26 Developer Challenge.

Tech highlights

Custom SQL parser and AST-driven query executor; hash-indexed JSON-backed persistent storage; supports CREATE TABLE, INSERT, SELECT (with JOIN and WHERE), UPDATE, and DELETE with type coercion and constraint checking; interactive REPL plus a Flask web demo.

Results

Shipped a working SQL engine with JOINs, indexing, and a REPL — built entirely on the Python standard library.

Challenges

Implementing JOINs, WHERE filtering, and type coercion on top of a hand-rolled hash index using only the Python standard library — no parser generator, no ORM — and keeping the JSON-backed storage consistent across crashes without a real write-ahead log.