Part 7. Logical Data Structures

Transcription

1 Part 7 Logical Data Structures

2 Logical Database Design Constructive approach Considers semantics Documents data dependencies identifiers entities needed relations rules Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 2

3 Logical Data Structures (LDS) Graphical means of naming and depicting the types of data in a database Simple, yet precise Useful to technically oriented analysts application-oriented users Easy to read Supports the design task logical structure design is hard tool aids the design task notation does not get in the way Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 3

4 Entity Basic LDS Components any type of thing about which information is maintained entity_name EXAMPLE student Attribute a characteristic of exactly one entity (fully functionally dependent on the entity) attribute_name EXAMPLE: Student attributes student student_name student_id# soc_sec# Relationships an association between a pair of entities (or roles ), one-to-one, one-to-many only or but never Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 4

5 Example Relationships 1-1 Example: Monogamous marriage man woman Can label relationship man wife of man/ husband of woman woman 1-M Example: Students of a college college student Need not label a relationship if it can be stated as: college of student / students of college or student has college / college has students Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 5

6 Handling an M-M Relationship M-M Example: Brother - Sister man_name man sisters of man/ brothers of woman woman woman_name Problem: how do you represent the presence of sibling rivalry? THIS WON'T WORK man_name woman_name man woman rivalry SOLUTION man_name woman_name man woman brother-sister rivalry Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 6

7 Identifier Representation Identifier: a set of attributes or relationships that uniquely identify an instance of an entity (single field key) (multiple-field key) Example: college_name student_name student_id# college student college# soc_sec# Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 7

8 Primary Key / Candidate Key state_name city_name state city state_abbrev city# primary key candidate key Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 8

9 Sample Database Employee: (emp) attributes: Ename, Job, Mgr, Hired, Rate, Bonus Department: (dept) attributes: DeptNo, Dname, Loc, Dbudget Task: (task) attributes: Tname, Hours Project: (proj) attributes: Project_id, Description, Pbudget, Due_date Relationships employees are members of a department employees have a manager who is an employee employees are assigned to tasks on projects Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 9

10 LDS for Sample Database DeptNo Dname Loc dept Ename Job Hired Rate Dbudget emp Mgr Bonus Tname task Project_id Description Hours proj Pbudget Due_date Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 10

11 Functional Dependency Revisited DeptNo Dname Loc Dbudget Ename Job Rate dept emp DeptNo identifies dept instances DeptNo --> Dbudget dependent on DeptNo DeptNo --> Loc on DeptNo Dname is an alternate key Dname --> Dbudget dependent on Dname Ename identifies emp instances Ename --> Job on Ename Ename --> Rate dependent on Ename Dbudget is fully functionally Loc is fully functionally dependent Dbudget is fully functionally Job is fully functionally dependent Rate is fully functionally an employee instance determines exactly one department Ename --> DeptNo DeptNo is fully functionally dependent on Ename Ename --> Loc Loc is fully functionally dependent on Ename, but this is a transitive full functional dependence Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 11

12 LDS for Example 1 - Suppliers A supplier supplies many parts, and a part can be supplied by many suppliers supplier part supp part_type availability Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 12

13 LDS for Example 2 - Inventory A product can be stored in many warehouses and a warehouse can contain many products part# warehouse# wh_address product warehouse inventory quantity Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 13

14 LDS for Example 3 - Departments A department can have many employees, and employee can only be in one department dept dept_loc name department employee Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 14

15 LDS for Example 4 - Locations Departments have one number, one name, and one location dept# dept_name dept_loc department Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 15

16 LDS for Example 5 - Stock An inventory is comprised of combinations of various parts from various suppliers - a supplier can supply many parts, and a part can be supplied by many suppliers p# s# sname part supplier inventory qty Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 16

17 LDS for Example 6 - Enrollment A student can take many subjects; a subject can be taken by many students. A subject can be taught by many teachers, a teacher can teach only one subject. A student can be taught by many teachers, a teacher can teach many students. student subject stu subj registration teach teacher Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 17

18 LDS for EXAMPLE 7 - SKILLS Employees can have many skills, and a skill can be had by many employees; an employee can know many languages and a language can be known by many employees. employee skill emp job_skill emp/lang/job_skill This diagram is correct if all 3 are interdependent language lang employee emp job_skill skill language This diagram is almost never correct lang (It implies that a skill can be held by only one employee) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 18

19 CORRECT LDS for INDEPENDENCE Assuming job skills and language skills are independent, they represent two separate many-to-many relationships employee emp/job_skill skill emp job_skill emp-lang language lang Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 19

20 LDS for EXAMPLE 8 - DEALERSHIPS In the general case, a contract involves one dealer, one manufacturer, and one product. A dealer can have many contracts, a manufacturer can have many contracts, and a product can be mentioned in many contracts. agent company dealership manufacturer contract This diagram is correct in the general case product vehicle Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 20

21 DEALERSHIPS with CONSTRAINTS Dealers can deal with many manufacturers, and manufacturers with many dealers. Dealers can sell many vehicle types and vehicle types can be sold by many dealers. Manufacturers can make many vehicles and vehicles can be made by many manufacturers. The combinations of who sells what is determined by symmetry. agent dealer-mfgr company dealer manufacturer dealer-vehicle mfgr-vehicle vehicle product Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 21

22 LDS for EXAMPLE 9 - CUSTOMERS A branch has many customers; a customer is in only one branch. There are only a limited number of legal branch names. branch cust# legal_branch customer Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 22

23 Entities: it must have identifier attributes relationships Modeling Concepts it must be the focus of the system need to develop for it : name description membership criteria must examine roles within subsets of it Attributes: must be non-transitively fully functionally dependent on the entity it describes must develop for each attribute: name description domain definition Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 23

24 Concept of Roles when 2 entities share a set of attributes OR when 2 entities have more than one relationship between them OR when subsets of an entity-instance have different attributes OR when subsets of an entity-instance participate in different relationships THEN MULTIPLE ROLES EXIST Examples of roles in the sample database: In emp, an employee plays 2 roles: works in a department (some) manages department In emp regular employees report to managers in the same department managers report to managers in a different department In emp, certain employees eligible for bonus (even if 0) ROLES ARE DOCUMENTED WHEN THEY ARE SIGNIFICANT Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 24

25 Alternate LDS for Sample Database DeptNo Dname Loc dept Ename Job Hired Rate Dbudget emp isa Mgr Tname Bonus task Project_id Description Hours proj Pbudget Due_date This has changed the rule about the group of employees for whom the bonus is applicable. Previously, analysts were technically eligible, even if none of them actually received a bonus. Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 25

26 Identifiers: Modeling Concepts (Continued) determine which attributes are part of it verify uniqueness establish not null requirements Relationships: establish degree 1-1 or 1-M entity on 1 side must be functionally dependent on entity on M side develop: - name - definition incorporate constraints, rules note referential integrity - (values of foreign key must exist in key field of another relation) - (e.g. in the emp relation, if an employee is listed as being in department 402, then in the dept relation there must contain a row with a key value of 402) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 26

27 Entity Analysis Oneness Sameness Categorization Identification Other Modeling Methods Object Abstraction (Smith & Smith) Objective: Intellectual Manageability Create hierarchies of abstraction along 2 dimensions: - aggregation (has / part of) - generalization (is / subtype) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 27

28 Object Extraction Examples Aggregation (has / part of) department task Tname Hours employee project Ename Job Project_id Pbudget Generalization (is / subtype) employee Ename Rate programmer clerk analyst supervisor language_skills type_speed bonus other_employee Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 28

29 LDS Map LDS to Well-Formed Relations Relational Model entity relation name attribute descriptor attribute single-valued relationship attribute (foreign key) descriptor multi-valued relationship nothing descriptor 1-1 relationship either or both relationship descriptors are attributes 1-M relationship relationship descriptor with degree 1 (on the M side) is an attribute Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 29

30 LDS fi Relations Examples Example: College students college_name student_name student_id# college student college# soc_sec# college (college#, college_name) student (student#, college#, student_name, soc_sec#) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 30

31 Sample Database Relations (See page 10 for the Sample Database LDS) dept (DeptNo, Dname, Loc, Dbudget) emp in emp (Ename, Job, Mgr, Hired, Rate, Bonus, DeptNo) proj (Project_id, Description, Pbudget, Due_date) task (Tname, Ename, Project_id, Hours Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 31

32 Relations for Example 1 - Suppliers supplier part supp part_type availability supp (supplier) part_type (part) availability (supplier, part) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 32

33 Relations for Example 2 - Inventory part# warehouse# wh_address product warehouse inventory quantity product (part#) warehouse (warehouse#, wh_address) inventory (part#, warehouse#, quantity) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 33

34 Relations for Example 3 - Departments dept dept_loc name department employee department (dept, dept_loc) employee (name, dept) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 34

35 Relations for Example 4 - Locations dept# dept_name dept_loc department department (dept#, dept_name, dept_loc) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 35

36 p# Relations for Example 5 - Stock s# sname part supplier inventory qty part (p#) supplier (s#, sname) inventory (p#, s#, qty) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 36

37 Relations for Example 6 - Enrollment student subject stu subj registration teach teacher stu (student) subj (subject) teach (teacher, subject) registration (student, teacher) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 37

38 Relations for Example 7 - Skills employee emp/job_skill skill emp job_skill emp-lang language lang emp (employee) job_skill (skill) lang (language) emp/job_skill (employee, skill) emp-lang (employee, language) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 38

39 Relations for Example 8 - Dealerships (See page 21 for Dealership LDS with symmetry restrictions) dealer (agent) manufacturer (company) vehicle (product) dealer-mfgr (agent, company) dealer-vehicle (agent, product) mfgr-vehicle (company, product) Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 39

40 Copyright Thomas P. Sturm Logical Data Structures Part 7, Page 40