Single User Relational Database

Database Systems · In-Memory RDBMS from Scratch
About

SURLY is a fully functional, single-user relational database management system built entirely from scratch. Designed as a lightweight, in-memory database, it features its own custom query language capable of parsing and executing both Data Definition Language (DDL) and Data Manipulation Language (DML) commands. The system reads plain-text command files through a custom-built lexical analyzer, interprets the statements, and executes complex database operations without relying on any external libraries, frameworks, or persistent disk storage.

Key Highlights
—Built a lexical analyzer from scratch to parse a custom SQL-like query language, routing commands through dedicated parser classes to execute in-memory operations.
—Developed robust logic for creating, destroying, inserting, and deleting relations with strict schema validation, type checking, and length constraints.
—Implemented complex relational algebra transformations (SELECT, PROJECT, JOIN) alongside a condition evaluation engine that handles AND/OR logic precedence and six relational operators.
—Architected the entire database system including data storage, a metadata registry (CATALOG), and a dynamically formatted tabular print system using exclusively the Java Standard Library with zero external dependencies.
Skills
JavaRelational AlgebraLexical AnalysisLinkedListHashSet
Images
Executing the SURLY interpreter from the command line. The system reads a plain-text command file, creates relation schemas (STUDENT, COURSE, ENROLLMENT), loads tuples via INSERT commands, and displays the STUDENT relation in a dynamically formatted table.

Executing the SURLY interpreter from the command line. The system reads a plain-text command file, creates relation schemas (STUDENT, COURSE, ENROLLMENT), loads tuples via INSERT commands, and displays the STUDENT relation in a dynamically formatted table.

Demonstrating the SELECT relational algebra operation with WHERE clause filtering. CS_STUDENTS filters by major, GOOD_STUDENTS uses a numeric comparison (GPA > 3.5), and SOPHOMORE_CS combines two conditions with AND logic — all producing temporary relations.

Demonstrating the SELECT relational algebra operation with WHERE clause filtering. CS_STUDENTS filters by major, GOOD_STUDENTS uses a numeric comparison (GPA > 3.5), and SOPHOMORE_CS combines two conditions with AND logic — all producing temporary relations.

The JOIN operation in action. STUDENT_COURSES pairs each student with their enrollments using a nested-loop theta join on matching IDs. FULL_ENROLLMENT chains a second join to bring in course titles and credits, demonstrating multi-table query composition.

The JOIN operation in action. STUDENT_COURSES pairs each student with their enrollments using a nested-loop theta join on matching IDs. FULL_ENROLLMENT chains a second join to bring in course titles and credits, demonstrating multi-table query composition.

Conditional tuple deletion using DELETE WHERE with compound conditions. The first command removes all enrollments with a grade of C, and the second targets a specific student-course pair using AND. Each operation reports the number of tuples removed and prints the updated relation.

Conditional tuple deletion using DELETE WHERE with compound conditions. The first command removes all enrollments with a grade of C, and the second targets a specific student-course pair using AND. Each operation reports the number of tuples removed and prints the updated relation.