School of Computer Science and Software Engineering. 2nd SEMESTER EXAMINATIONS 2007 CITS3240 DATABASES

Transcription

1 School of Computer Science and Software Engineering 2nd SEMESTER EXAMINATIONS 2007 SURNAME: GIVEN NAMES: STUDENT NO: SIGNATURE: This paper contains: 9 pages (including the title page) Time allowed: 2 hours 10 minutes Section A: 1 question Section B: 2 questions Section C: 5 questions (choose 4) TOTAL MARKS: 30 marks 30 marks 40 marks 100 marks You should attempt Section A, Section B and three questions in Section C. The questions should be answered in the supplied answer booklet. PLEASE NOTE Examination candidates may only bring authorised materials into the examination room. If a supervisor finds, during the examination, that you have unauthorised material, in whatever form, in the vicinity of your desk or on your person, whether in the examination room or the toilets or en route to/from the toilets, the matter will be reported to the head of school and disciplinary action will normally be taken against you. This action may result in your being deprived of any credit for this examination or even, in some cases, for the whole unit. This will apply regardless of whether the material has been used at the time it is found. Therefore, any candidate who has brought any unauthorised material whatsoever into the examination room should declare it to the supervisor immediately. Candidates who are uncertain whether any material is authorised should ask the supervisor for clarification.

2 2 Instructions This paper contains three sections totalling 100 marks. Section A is worth 40 marks, Section B is worth 30 marks and Section C is worth 30 marks. Candidates should attempt Section A, Section B and FOUR of the five questions in Section C. If more than four questions are attempted in Section C only the first four will be marked. If you change your mind part way through an answer you should clearly cross the answer out and write NOT ATTEMPTED Please fill in the question numbers in the spaces on the front of the answer booklet.

3 3 Section A 1. Consider a MySQL database with three tables created as follows: CREATE TABLE part ( pid INT NOT NULL AUTO_INCREMENT, description VARCHAR(128), PRIMARY KEY (pid) ) Engine = InnoDB; CREATE TABLE warehouse ( wid INT NOT NULL AUTO_INCREMENT, address VARCHAR(128), PRIMARY KEY (wid) ) Engine = InnoDB; CREATE TABLE stock ( pid INT, wid INT, quantity INT, unitcost FLOAT, FOREIGN KEY (pid) REFERENCES part (pid), FOREIGN KEY (wid) REFERENCES warehouse (wid) ) Engine = InnoDB; These tables contain information about various machinery parts stored in different warehouse locations. The stock table indicates the number of items currently held at a particular warehouse, and the cost of each item if it were to be delivered from that warehouse. For example, suppose that the table contained the following rows: pid wid quantity unitcost This would indicate Warehouse 1 has 10 units in stock of Part #1 and each one costs $23.50, while Warehouse 2 has 15 units in stock at a cost of $22.95 each, etc.

4 4 Suppose that you frequently need to answer questions of the form What is the cheapest total cost of an order of 23 units of Part #1? where the number of units and the id of the part change from query to query. In this case the cheapest way of supplying this order would be to purchase all 7 items in stock at Warehouse 3, all 15 from Warehouse 2 and just 1 item from Warehouse 1 at a total cost of = (a) Write a Java method public double totalcost(int pid, int numitems) that computes the cheapest total cost of supplying the specified number of units of part with id pid. You may assume that the method has access to a variable conn that refers to an already-created java.sql.connection. To simplify your method, you may assume that there is enough stock available in total to fulfill the request. (b) Write a MySQL stored procedure totalcost that takes two IN parameters pid and numitems and has a single OUT parameter. This procedure should calculate the cheapest total cost of fulfilling this order and assign this value to the OUT parameter. Again, you may assume that there is enough stock available in total. (15) (15)

5 5 Section B 2. Queries in SQL Consider the following schema which represents students, lecturers, units and enrolments in a university. Student(snum:int, sname:string) Lecturer(lnum:int, lname:string) Unit(uid:int, ucode:string, year:int, lnum:string) Enrolled(snum:int, uid:int, grade:int) The meanings of the fields are as follows: Field Meaning Example Student.snum Student Number Student.sname Name Gillian Tan Lecturer.lnum Staff Number Lecturer.lname Name William Biggs Unit.uid Artificial key 57 Unit.ucode Unit code Unit.year Year 2006 Unit.lnum Foreign key to Lecturer.lnum Enrolled.snum Foreign key to Student.snum Enrolled.uid Foreign key to Unit.uid Enrolled.grade Mark attained 67 Write MySQL queries for the following tasks: (a) List the student numbers and names of all students taking in (b) List the unit details (code, year and lecturer s name) for all units with fewer than 10 students enrolled. (c) For every student and year, list the number of units taken by the student in that year, and the average mark they achieved that year. (Do not attempt to deal with years where the student was not enrolled.)

6 6 (d) List the top student in for each year from (there may be more than one top student in any given year if they get the same mark.) (e) List the students who have obtained HDs (that is, a mark of 80% or more) in every one of their completed units. 3. ER diagramming Draw an ER diagram (using the symbols and conventions outlined in lectures) that captures as far as possible the following requirements for a database for a car parts retailer, using natural keys where appropriate. (a) Each car part has a unique part number and description. (b) Parts manufacturers have a unique name and an address and phone number. (c) Each manufacturer has a fixed price for each part, but the same part may be produced at different prices by a number of different manufacturers. (d) The database stores customer details including name and delivery address. (e) Customers place orders that may contain one or more items, where each item consists of a quantity of a particular brand of part (e.g 4 x 215/R16 tyres from Dunlop). (f) Each order has a dispatch date associated with it. (g) Every part in the catalogue is manufactured by at least one manufacturer and every order contains at least one item. Make sure you incorporate any participation and key constraints implied by the specifications. (Draw a single ER-diagram that includes all the entities and relationships in one big picture, rather than splitting it into a number of small diagrams.) (15)

7 7 Section C 4. Database Applications (a) Give a brief outline of how a database abstraction layer such as PHP s MDB2 works, and explain the advantages and disadvantages of using a database abstraction layer. (b) What are the factors that a DBA should consider when considering whether to use stored procedures and/or client programs (e.g. Java or PHP) in a database application? (6) (4) 5. Data Mining Consider the following data for market-basket analysis where each line contains the items for a single transaction. Apples, Doughnuts, Eggplant, Fava Beans Doughnuts, Eggplant, Fava Beans Apples, Fava Beans Banana, Candy, Eggplant, Fava Beans Eggplant, Fava Beans Banana, Candy, Doughnuts, Eggplant Banana, Doughnuts, Fava Beans Doughnuts, Eggplant, Fava Beans Doughnuts, Fava Beans Eggplant, Fava Beans (a) Find the frequent itemsets with minimum support 50%. (5) (b) Find all association rules with minimum support 50% and minimum confidence 50%. (5)

8 8 6. Normalization A company assigns its employees to various client projects and keeps track of the hours worked on each particular project by each employee. The client is then billed for those hours at an hourly rate that depends on the employee s level. Currently it keeps this information in a flat file format as shown below: ProjId ProjName EmpId EmpName EmpLevel Rate Hours 15 Evergreen 101 John Barnes Evergreen 105 Jill Bates Evergreen 110 James Callan SilverMoon 110 James Callan SilverMoon 115 Alice Tan Use the theory of database normalization to design a well-structured relational database schema to store data of this type, epxlaining your reasoning. (10) 7. Transactions Isolation is the property that each database user can complete transactions as though they are the only user of the database. (a) What is interleaving? (2) (b) Name and describe (with examples) the three interleaving anomalies that can violate the property of isolation. (6) (c) Why does a DBMS use interleaving if it can cause these anomalies? (2)

9 9 8. Relational Algebra Consider the following schema which models university lecturers working on various research projects for a certain percentage of their time. lecturer (lid:int, name:string) project (pid:int, agency:string, budget:real) works(lid:int, pid:int, pc:real) You may assume that L, P and W are aliases for lecturer, project and works. (a) Write a relational algebra query that lists the names of all lecturers that work for more than 10% of their time on any ARC project with a budget of more than $ (b) Write a relational algebra query that lists the names of all lecturers who work only on ARC projects. (c) Write a relational algebra query to list the names of the lecturers who work on every ARC project. (4)