Databases
Relational databases: ER diagrams, relational model, SQL, schema design, normalization, functional dependencies, concurrency.
Exercise attendance is not mandatory but is strongly recommended. Register for both an A and a B exercise group in Canvas; the homework itself is individual.
Learning objectives
The students will learn: to design and evaluate database schemas using Entity-Relationship diagrams, (Applying knowledge and understanding) (Making judgements) to understand the relational model (including integrity constraints such as key constraints and foreign key constraints), (Knowledge and understanding) to translate Entity-Relationship diagrams into relational schemas (and to enforce cardinality constraints by design), (Applying knowledge and understanding) to reason about good and bad database design using functional dependencies, (Making judgements) to apply normalization algorithms to transform a database schema into (various) normal forms, (Applying knowledge and understanding) to understand concurrency concepts like transactions, schedules and concurrency anomalies, and (Knowledge and understanding) to reason about different techniques for concurrency control (lock based concurrency control, optimistic concurrency control, multiversion concurrency control). (Making judgements)
Databases teaches the complete relational-database workflow: modelling a domain with ER diagrams, translating that model into a relational schema, querying and updating it with SQL, improving schema quality with functional dependencies and normal forms, and reasoning about transactions and concurrent schedules. Database APIs and safe parameterised queries are also examinable.
How the course runs. There are no traditional lectures. You prepare from short lecture videos and the textbook in your own time, and the contact hours are two weekly question hours plus roughly two on-campus exercise classes per week. You sign up on Canvas for both an A group (which works through that week's exercise sheet) and a B group (which is time to work on the homework with TAs available). The last A class of the block goes through the example exam.
The homework is a gate, not just points. All homework runs in iSubmit and is individual, with the single exception of the conceptual modelling assignment, which may be done in pairs. You need 70% of the homework points to qualify for the exam or the resit, and qualification does not carry over from a previous year. Hit 90% and you get a +0.5 bonus on a passing grade.
The exam. Closed book, on campus, and taken in iSubmit rather than on paper. The task types mirror the homework: draw an E/R diagram, write SQL, apply the synthesis algorithm, build a precedence graph, plus multiple-choice and open theory questions. Most tasks give partial credit and immediate feedback, and some automatically graded ones limit your attempts.
Worth knowing before you sign up. The workload is steady rather than spiky, but the homework deadlines run through the whole block, so this is a bad course to fall behind in. If you liked the structured, rule-following parts of Logic and Sets, you will probably enjoy this one.
Assessment
Individual homework is mandatory: in 2026 you need at least 55 homework points (70%) to qualify for the exam or resit. A score of at least 75 points (90%) earns a +0.5 bonus, applied only after passing the exam with at least 5.5. Homework eligibility from a previous course year does not carry over. The final exam is a closed-book, on-campus digital exam in iSubmit with practical tasks similar to the homework plus multiple-choice and open theory questions; some automatically graded tasks have limited attempts.
Teaching methods
Self-study lecture videos and textbook reading, two weekly question hours, approximately two on-campus exercise classes per week, and mandatory individual homework in iSubmit.
Literature
Database Systems, The Complete Book, by: Hector Garcia-Molina & Jeffrey D. Ullman & Jennifer Widom. 2nd edition.