INFO90002 Database Systems and Data Modelling
The Complete Exam Bible · S2 2026INFO90002 Overview
- The University of Melbourne
- Semester 2, 2026
- Postgraduate coursework, Masters level (AQF level 9)
Within The University of Melbourne — Faculty of Engineering and Information Technology (FEIT), School of Computing & Information Systems (CIS), INFO90002 takes a set of business rules all the way down the database design lifecycle — conceptual ER modelling in Chen and Crow's Foot notation, then logical and physical design and normalisation, then SQL DDL/DML on a live MySQL server (joins, functions, subqueries, views), and finally the 'advanced' half: transactions, concurrency and locking, administration and architecture, distributed databases, data warehousing, NoSQL/JSON, and database security and backup.
- INFO90002 grading 20% group data-modelling assignment (teams of 4 from the same tute, due end of Week 6), 10% individual SQL assignment (due end of Week 9), 15% closed-book invigilated Quiz 1 (Week 7 lecture), 5% open-book online Quiz 2 (Week 12 lecture, first hour), 50% end-of-semester exam
- INFO90002 task mode 2 hours, November, on campus / in person, closed book and invigilated; dictionaries are not allowed at the exam. Quiz 1 is also closed book, sat in the Week 7 lecture at 5.30 pm in Kwong Lee Dow via Lockdown Browser; Quiz 2 is online and open book. (Note a source conflict to resolve before publishing: the syllabus week table places Quiz 1 in Week 6, while the home page and the admin deck both place it in Week 7 with a date of 8 September — use Week 7.)
- INFO90002 mark trap Normalisation and the advanced-theory half, plus instruction-following on the assignments. The coordinator's own slide: Industry students generally do poorly because: • Don't follow the instructions • Don't do well in normalisation and advanced theory topics. The second named failure mode is GenAI dependence: My experience is that you repeat GenAI mistakes in the exam (and lose marks) … If GenAI did your assignments, only a miracle will make you pass the exam • Has not happened so far.
- INFO90002 pass rule Two independent hurdles, stated quantitatively on the admin deck: Hurdle 1 A1+A2+Q1+Q2 must be >=25% and Hurdle 2 Exam must be >=25%. The home page states the same rule qualitatively: This subject has hurdle requirements. To pass INFO90002, you must pass the Assignments and Quizzes together (hurdle 1), AND the End of Semester Assessment (Exam, hurdle 2).
How INFO90002 is assessed
| Component | Weight | Format |
|---|---|---|
| Assignment 1: Data Modelling | 20% | Group assignment |
| Assignment 2: SQL | 10% | Individual assignment |
| Quiz 1 | 15% | Individual timed closed-book quiz in the retrieved notice |
| Quiz 2 | 5% | Individual timed open-book quiz in the retrieved notice |
| End-of-Semester Examination | 50% | Two-hour on-campus examination in the retrieved notice |
| Coursework Hurdle · hurdle | 25/50 minimum | Combined A1, A2, Quiz 1 and Quiz 2 |
| Examination Hurdle · hurdle | 25/50 minimum | End-of-semester examination |
The five components total 100%. Hurdle 1 requires at least 25 marks from the 50-mark coursework group (A1, A2 and both quizzes); Hurdle 2 requires at least 25 marks from the 50-mark exam. Later canonical explanations are not presented as University questions or unpublished task content.
What INFO90002 covers
The learning path moves from Assessment, Evidence and Database Lifecycle, through the problems opened by SQL Aggregation, Subqueries and Views, to the synthesis required in Security, Backup and Integrated Exam Reasoning.
Assessment, Evidence and Database Lifecycle
20/10/15/5/50 structure · database lifecycle · two-hurdle control · map requirements, design, implementation and validation to the assessed workflow02Requirements and Conceptual Modelling
business rules · entities and attributes · relationships · translate stakeholder language into testable data requirements03ER Diagrams and Cardinality
keys and identifiers · cardinality and participation · associative entities · justify every relationship constraint with a business rule04Logical Design and Normalisation
functional dependency · normal forms · lossless decomposition · remove update anomalies while preserving meaning and joinability05Relational Algebra and SQL Foundations
selection and projection · joins · set and bag behaviour · predict a query result from relational operations before running SQL06SQL DDL, DML and Integrity
table definition · constraints · insert update delete · encode business rules at the strongest appropriate database boundary07SQL Aggregation, Subqueries and Views
grouping and HAVING · nested query · view abstraction · choose a query shape that preserves the intended row and group grain08Transactions, Concurrency and Recovery
ACID properties · locking and isolation · logging and recovery · trace an interleaving to the anomaly and isolation control09Architecture, Administration and Distribution
three-tier architecture · catalogue and administration · distributed data · place responsibility across client, service and database layers10Data Warehousing and Analytical Models
operational versus analytical data · fact and dimension · ETL and lineage · design an analytical grain and preserve the transformation trail11NoSQL, JSON and Semi-Structured Data
document model · schema flexibility · consistency trade-off · select a data model from access pattern, integrity and evolution needs12Security, Backup and Integrated Exam Reasoning
least privilege · backup and restore · integrated design trade-off · combine modelling, SQL and operational controls in one defensible database decisionIt is Postgraduate coursework, Masters level (AQF level 9).
It is a build-it-then-defend-it subject: coursework is 30% of artefacts you actually produce (a group Crow's Foot data model in draw.io/MySQL Workbench, and an individual SQL script that must run on the university's own info90002db.eng.unimelb.edu.au server), but 70% of the mark comes from timed assessment, and the grade is gated by two independent 25/50 hurdles — coursework and exam must each be passed on their own.
Both Quiz 1 and the exam are closed-book and invigilated (Quiz 1 via Lockdown Browser), and the coordinator states outright that industry-experienced students who already know SQL still lose marks because they 'don't follow the instructions' and 'don't do well in normalisation and advanced theory topics'.
Assessment in INFO90002 is distributed as follows: 20% group data-modelling assignment (teams of 4 from the same tute, due end of Week 6), 10% individual SQL assignment (due end of Week 9), 15% closed-book invigilated Quiz 1 (Week 7 lecture), 5% open-book online Quiz 2 (Week 12 lecture, first hour), 50% end-of-semester exam
The operational assessment conditions matter here.
2 hours, November, on campus / in person, closed book and invigilated; dictionaries are not allowed at the exam. Quiz 1 is also closed book, sat in the Week 7 lecture at 5.30 pm in Kwong Lee Dow via Lockdown Browser; Quiz 2 is online and open book.
(Note a source conflict to resolve before publishing: the syllabus week table places Quiz 1 in Week 6, while the home page and the admin deck both place it in Week 7 with a date of 8 September — use Week 7.)
What makes INFO90002 demanding is concrete: Normalisation and the advanced-theory half, plus instruction-following on the assignments.
The coordinator's own slide: Industry students generally do poorly because: • Don't follow the instructions • Don't do well in normalisation and advanced theory topics.
The second named failure mode is GenAI dependence: My experience is that you repeat GenAI mistakes in the exam (and lose marks) … If GenAI did your assignments, only a miracle will make you pass the exam • Has not happened so far.
Two independent hurdles, stated quantitatively on the admin deck: Hurdle 1 A1+A2+Q1+Q2 must be >=25% and Hurdle 2 Exam must be >=25%.
The home page states the same rule qualitatively: This subject has hurdle requirements.
To pass INFO90002, you must pass the Assignments and Quizzes together (hurdle 1), AND the End of Semester Assessment (Exam, hurdle 2).
The learning path moves from Assessment, Evidence and Database Lifecycle, through the problems opened by SQL Aggregation, Subqueries and Views, to the synthesis required in Security, Backup and Integrated Exam Reasoning.
Worked example: Database Systems and Data Modelling integrated response
- 1Extract the outcome, actor or operation that the Database Systems and Data Modelling integrated response task actually requires.
- 1State the precondition under which three-tier architecture is relevant rather than merely familiar.
- 1Use selection and projection to reject the nearest alternative, then run a failure-path check with ACID properties.
- 1Choose the response and state when it must be withdrawn or narrowed: Serialisable behaviour can carry performance and deadlock costs.
Key terms
- conceptual, logical and physical design (the database development lifecycle)
- Conceptual design models business entities and relationships independently of technology, logical design translates them into a data model and constraints, and physical design specifies storage, indexes and implementation details.
- Crow's Foot vs Chen notation
- Crow's Foot notation shows cardinality with line-end symbols around entity boxes, whereas Chen notation represents entities, attributes and relationships with rectangles, ovals and diamonds.
- connectivity vs cardinality (and optional/mandatory participation)
- Connectivity names the relationship type such as one-to-many, cardinality states minimum and maximum participation counts, and optional or mandatory participation indicates whether the minimum is zero or one.
- identifier / primary key; composite, derived and multi-valued attributes
- An identifier uniquely distinguishes entity occurrences and becomes a primary key in a relation; a composite attribute has components, a derived attribute is calculated, and a multi-valued attribute can hold several values per entity.
- normalisation
- Normalisation decomposes relations according to functional dependencies to reduce redundancy and prevent insertion, update and deletion anomalies while preserving required information.
- DDL, DML and DCL (CREATE/DROP/ALTER vs SELECT/INSERT/UPDATE/DELETE vs GRANT/REVOKE)
- DDL defines database structures with commands such as CREATE, ALTER and DROP; DML queries or changes data with SELECT, INSERT, UPDATE and DELETE; DCL manages privileges with GRANT and REVOKE.
- transactions, concurrency and locking
- A transaction is a logical unit of database work that should satisfy ACID properties; concurrency control and locking coordinate simultaneous transactions to prevent inconsistent or lost updates.
- referential integrity; logical vs physical data independence
- Referential integrity requires each foreign key to match an existing referenced key or be null when allowed; logical and physical data independence protect users from changes to schemas or storage respectively.
INFO90002 FAQ
Is INFO90002 hard?
Normalisation and the advanced-theory half, plus instruction-following on the assignments. The coordinator's own slide: Industry students generally do poorly because: • Don't follow the instructions • Don't do well in normalisation and advanced theory topics.
The second named failure mode is GenAI dependence: My experience is that you repeat GenAI mistakes in the exam (and lose marks) … If GenAI did your assignments, only a miracle will make you pass the exam • Has not happened so far.
How is INFO90002 assessed?
20% group data-modelling assignment (teams of 4 from the same tute, due end of Week 6), 10% individual SQL assignment (due end of Week 9), 15% closed-book invigilated Quiz 1 (Week 7 lecture), 5% open-book online Quiz 2 (Week 12 lecture, first hour), 50% end-of-semester exam
What is the INFO90002 exam or final-task format?
2 hours, November, on campus / in person, closed book and invigilated; dictionaries are not allowed at the exam. Quiz 1 is also closed book, sat in the Week 7 lecture at 5.30 pm in Kwong Lee Dow via Lockdown Browser; Quiz 2 is online and open book.
(Note a source conflict to resolve before publishing: the syllabus week table places Quiz 1 in Week 6, while the home page and the admin deck both place it in Week 7 with a date of 8 September — use Week 7.)
Does INFO90002 have a hurdle or component-level pass rule?
Two independent hurdles, stated quantitatively on the admin deck: Hurdle 1 A1+A2+Q1+Q2 must be >=25% and Hurdle 2 Exam must be >=25%. The home page states the same rule qualitatively: This subject has hurdle requirements. To pass INFO90002, you must pass the Assignments and Quizzes together (hurdle 1), AND the End of Semester Assessment (Exam, hurdle 2).
What prerequisites or restrictions apply to INFO90002?
Check the current official handbook before enrolling in INFO90002; prerequisites are not inferred from its course number.
Is INFO90002 offered in Semester 2, 2026?
This resource is aligned to Semester 2, 2026. Confirm your class and assessment timetable in the current institutional system.
Is this INFO90002 resource an official university guide?
No. It is an independent INFO90002 study resource; current institutional instructions remain authoritative for assessment operation.
How to study for the exam
Retrieve the subject map, practise the recurring method—move from user requirements to a constrained data model, implement it in SQL and test integrity, concurrency and security consequences—on changed scenarios, and verify every operational assessment detail in the live institutional system.
Your AI Information Technology tutor for INFO90002
Stuck on a hard INFO90002 question? Sia is AskSia’s AI Information Technology tutor — ask any INFO90002 Database Systems and Data Modelling question and get a clear, step-by-step explanation grounded in how the course is actually taught and assessed. Read this whole study guide free, then take your hardest questions to Sia.
