Learning SQL from within: integrating database exercises into the database itself

TL;DR

This paper introduces SQLab, an innovative framework embedding SQL exercises within the database itself, using query fingerprinting for enhanced feedback and enabling complex, storyline-driven game-based learning for SQL education.

Contribution

It presents a novel query fingerprinting technique for evaluating SQL answers and demonstrates its application in an open-source, cross-vendor framework for interactive SQL learning.

Findings

Successful implementation of SQLab with 30 exercises and 2 games.

Positive student engagement over a three-year testing period.

Enhanced feedback mechanism through query fingerprinting.

Abstract

SQL adventure builder (SQLab) is an open-source framework for creating SQL games that are embedded within the very database they query. Students' answers are evaluated using query fingerprinting, a novel technique that allows for better feedback than traditional SQL online judge systems. Fingerprints act as tokens that are used to unlock messages encrypted in an isolated auxiliary table. These messages may include hints, answer keys, examples, explanations, or narrative elements. They can also contain the problem statement of the next task, which turns them into nodes in a virtual DAG with queries as edges. This makes it possible to design a coherent adventure with a storyline of arbitrary complexity. This paper describes the theoretical underpinnings of SQLab's query fingerprinting model, its implementation challenges, and its potential to improve SQL education through game-based learning. The underlying concepts are fully cross-vendor, and support for SQLite, PostgreSQL and MySQL is already available. As a proof of concept, two games, 30 exercises and one mock exam were tested over a three-year period with about 300 students.