Omar Abdul Baki ID: Database Systems Assignment 1 January 27, Question 1.

Transcription

1 Question 1. Omar Abdul Baki ID: Database Systems Assignment 1 January 27, Please refer to the attached diagram 1 2. Please refer to the attached diagram Doctor(SIN: string, name: string, specialty:string, yearsexeperience:integer) CREATE TABLE Doctor ( SIN CHAR(9), name CHAR(20), specialty CHAR(20), yearsexperience INTEGER, PRIMARY KEY(SIN)) Rationale: 1. SIN is chosen as the primary key because this field uniquely identifies a doctor. Sees(PSIN: string, DSIN: string) CREATE TABLE Sees ( PSIN CHAR(9), DSIN CHAR(9), PRIMARY KEY(DSIN, PSIN)) FOREIGN KEY (PSIN) REFERENCES Patient(SIN), FOREIGN KEY (DSIN) REFERENCES Doctor(SIN)) Rationale: 1. DSIN and PSIN form the primary key since they are also primary key attributes of the entities that are associated with this relationship. In this relationship, participation must be total for a doctor since all doctors must have at least 1 patient. Moreover, participation must be also total for patient since all patients have at least one primary physician. Patient( SIN: string, name:string, address:string, age:integer, primaryphysician:string) SQL Schema CREATE TABLE Patient ( SIN CHAR(9), name CHAR(20), address CHAR(50), age INTEGER primaryphysician CHAR(9) NOT NULL, FOREIGN KEY (primaryphysician) REFERENCES Doctor(SIN), PRIMARY KEY(SIN)) Rationale: 1. SIN is chosen as the primary key because this field uniquely identifies a patient. 2. primaryphysician is an attribute of Patient since the PrimaryDoctor(Doctor-Patient relationship is many to 1. Moreover this field is Not Null since every patient must have a - 1 -

2 primary physician. Prescribes(PSIN:string, DSIN:string, pharmaceuticalname:string, drugname:string, prescribedate:date, quantity:integer) SQL Schema CREATE TABLE Prescribes ( PSIN CHAR(9), DSIN CHAR(9), pharmaceuticalname CHAR(20), drugname CHAR(20), prescribedate DATE, quantity INTEGER, PRIMARY KEY (PSIN, DSIN, drugname, pharmaceuticalname), FOREIGN KEY (PSIN) REFERENCES Patient(SIN), FOREIGN KEY (DSIN) REFERENCES Doctor(SIN), FOREIGN KEY (drugname) REFERENCES Drug(tradeName), FOREIGN KEY (pharmaceuticalname) REFERENCES Pharmaceutical(name)) Rationale: 1. PSIN, DSIN, drugname and pharmaceuticalname collectively constitute the Prescribes primary key. These fields are primary key fields since they are also primary key attributes of the entities that are associated with this relationship. Moreover the requirements state that if a doctor prescribes the same drug for the same patient more than once, then only the last such prescription needs to be stored. Thus other attributes of the Prescribes relation need not be included in the primary key. SQL Schema Drug (tradename:string, formula:string, suppliername:string) Note: tradename is a discriminator. The tradename and the suppliername foreign key together make up the Drug primary key. CREATE TABLE Drug ( tradename CHAR(20), formula CHAR(50), suppliername CHAR(20), PRIMARY KEY (tradename, suppliername), FOREIGN KEY (suppliername) REFERENCES Pharmaceutical(name) ON DELETE CASCADE) Rationale: 1. This relation is a weak entity since a Drug is uniquely identified using both its tradename and the suppliername(a primary key of the pharmaceutical entity). 2. The primary key constitutes the discrimator (tradename) and the suppliername foreignkey. 3. The Cascade statement ensures that upon deletion of a pharmaceutical tuple, its - 2 -

3 associated drugs are deleted. This is derived from the requirement stating: If a pharmaceutical company is deleted, we need not keep track of its products any longer. Pharmaceutical(name:string, phonenumber:string) SQL Schema CREATE TABLE Pharmaceutical ( name CHAR(20), phonenumber CHAR(10), PRIMARY KEY (name)) Rationale: 1. the pharmaceutical company name is chosen as the primary key because this field uniquely identifies the company. SQL Schema Pharmacy(name:string, address:string, phonenumber:string) CREATE TABLE Pharmacy ( name CHAR(20), address CHAR(50), phonenumber CHAR(10), PRIMARY KEY (name, address, phonenumber)) Rationale: 1. the pharmacy name, address, and phonenumber are collectively chosen as the primary key because no requirement states that any of these fields are unique. SQL Schema Sells( pharmacyname:string, pharmacyaddress:string, pharmacyphonenumber:string, drugtradename:string, pharmaceuticalname:string, price:real) CREATE TABLE Sells ( pharmacyname CHAR(20), pharmacyaddress CHAR(50), pharmacyphonenumber CHAR(10), drugtradename CHAR(20) pharmaceuticalname CHAR(20), price REAL, PRIMARY KEY (pharmacyname, pharmacyaddress, pharmacyphonenumber, drugtradename, pharmaceuticalname) FOREIGN KEY (pharmacyname) REFERENCES Pharmacy(name), FOREIGN KEY (pharmacyaddress) REFERENCES Pharmacy (address), FOREIGN KEY (pharmacyphonenumber) REFERENCES Pharmacy(phoneNumber), FOREIGN KEY (drugtradename) REFERENCES Drug(tradeName), FOREIGN KEY (pharmaceuticalname) REFERENCES Pharmaceutical(name)) - 3 -

4 Rationale: 1. pharmacyname, pharmacyaddress, pharmacyphonenumber, drugtradename and pharmaceuticalname collectively constitute the Sells primary key. These fields are primary key fields since they are also primary key attributes of the Drug and Pharmacy entities associated with this relationship. SQL Schema Contracts(pharmacyName:string, pharmacyaddress:string, pharmacyphonenumber:string, pharmaceuticalname:string, pharmaceuticalphonenumber:string, supervisor, startdate:date, enddate:date, text:string) CREATE TABLE Contracts ( pharmacyname CHAR(20), pharmacyaddress CHAR(50), pharmacyphonenumber CHAR(10), pharmaceuticalname CHAR(20), supervisor CHAR(20) NOT NULL, startdate DATE, enddate DATE, text CHAR(10000) PRIMARY KEY (pharmacyname, pharmacyaddress, pharmacyphonenumber, pharmaceuticalname, text ) FOREIGN KEY (pharmacyname) REFERENCES Pharmacy(Name), FOREIGN KEY (pharmacyaddress) REFERENCES Pharmacy (address), FOREIGN KEY (pharmacyphonenumber) REFERENCES Pharmacy(phoneNumber), FOREIGN KEY (pharmaceuticalname) REFERENCES Drug(name) ) Rationale: 1. pharmacyname, pharmacyaddress, pharmacyphonenumber, pharmaceuticalname, and text collectively constitute the Contracts primary key. These fields are primary key fields since they are also primary key attributes of the Drug and Pharmaceutical entities associated with this relationship. Text is also included in the primary key since we assume it is possible for a pharmacy to have multiple contracts with the same pharmaceutical company. Hence the text field allows us to distinguish between different contracts between the same pharmacy and pharmaceutical companies. 2. The Not Null constraint on supervisor ensures that a contract always has a supervisor as stated in the requirements

5 4. If drugs were to be sold at a fixed price by all pharmacies then we can simplify our design by removing price attribute from the sells relationship between Pharmacy and drugs. Since the price is now globally constant amongst each drug at all pharmacies, the price can effectively become an attribute that can be added to the drug entity in both the ER and Relational Database schemas. Modified Database Schema: Figure 1-Modified Entity Relation Model Drug (tradename, suppliername, formula, price) Sells( pharmacyname, pharmacyaddress, pharmacyphonenumber, drugtradename, pharmaceuticalname) Question An example of a foreign key constraint involving the Department relation is that the did field in the Works relation is a foreign key which references the primary key in the Departments relation. The options for enforcing this constraint are the following: a. NO ACTION: If the Works relation contains a record corresponding to the department being deleted, then disallow the deletion. This implies that a department can not be deleted as long as there are employees working in it. The records of employees working for that department must be deleted prior to deleting the department itself. b. CASCADE: This option allows the deletion of the department. Moreover, it will enforce that all tuples in the Works relation which reference that department will also be deleted. This implies that deleting a department will result in the deletion of all stored employee information for that department. c. SET DEFAULT: This option allows the deletion of the department. However, all tuples in the Works relation referencing the department to be deleted will now reference some predefined default department. c. SET NULL: This option allows the deletion of the department. However, all tuples in the Works relation referencing the department to be deleted will now have a NULL value in the did field. This means that deleting a department can result in records showing employees not working for any department

6 2. a. Get the id s and names of employees who have a salary of more than 50 k., ename( σ salary> 50000( Employee)) b. Get details on employees who work in more than one department. ρ ( Works2, Works) Temp1, Temp ( Works2 1 Temp 2. = Emloyee. 2. = Works. Works)) Works Works. did! = Works2. did Employee c. Who are the managers who work also in a department that they do not manage. Managers, Employee, ename, did ( Emloyee. = Department. manager _ id Department)), ename ( Managers Managers. = Works ) works. Managers. did! = Works. did - 6 -

7 d. Get the id s of employees who work in department 203 and also in department 302. ρ ( EmployeesA, σ Department did= 203 ) ρ ( EmployeesB, σ Department did = 302 ) ( EmployeesA. EmployeesB EmployeesA. = EmployeesB) E. Find those employees who do not work in the department managed by John Doe. JohnDoe, ( σ ename= ' JohnDoe ' JDDepartment, ( JohnDoe ( Employee))) did JohnDoe = Deparmtnet. manager _ id NonJDEmployees, ( JDDepartment. Department)) JDDeparment = Works. did. did! Works) - 7 -

8 Step 4: NonJDEmplo yees Employees F. Find details on the employee(s) who work(s) the maximum number of hours per week in department 101. Employees101, 2 ( σ did = 101 hours _ per _ week Employees hours1,4 ( Works)) Pr oduct(1 hours2), Employees101 Employees101 NonMaxHour s,, hours1( σ hours 2 hour1( Employees Pr oduct)) > MaxHour, Employees101 Step 4: NonMaxHour s) - 8 -

9 ( Employee MaxHour Employee = ename eage esalary MaxHour.. ) Question Get details on patients who have consulted Dr. Lee at least once. DrLee, ( σ did dname= ' Lee' ( Doctor ))) DrLeePatie nts, ( DrLee pid Consults DrLee did. did = DrLeePatients Patient. Consults )) 2. List those doctors id who do not work in Victoria (clinic name) for all clinics of Montreal. OtherClini cs, cid ( σ! = ' Victoria ' ( Clinic))) cname city = montreal - 9 -

10 ( OtherClini cs = Works _ in)) did Doctor OtherClini cs. cid Works _ in. cid 3. Find those patients name who have never been to a clinic in a city other than the one he/she lives in. ConsultCit ies,, city ( Clinic Clinic. cid Consults ))) pid = Consults. cid TravelledPatients, ConsultCit ies ConsultCities Patient. city ConsultCities. pid = Patient. pid Patients PatientTravelled. city! = Patient )