Agent 7 which languages? skills?

Transcription

1 Agent 7 which languages? skills? select * from languagerel where agent_id = 7 lang_id agent_id agent 7 speaks 4 languages select * from skillrel where agent_id = 7 skill_id agent_id agent 7 has 2 skills CS386/586 Introduction to Database Systems, Lois Delcambre Slide 1

2 Intersect vs. Intersect ALL SELECT agent_id FROM languagerel WHERE agent_id = 7 INTERSECT SELECT agent_id FROM skillrel WHERE agent_id = 7 agent_id Agent 7 has (at least one) language and (at least one) skill 7 SELECT agent_id FROM languagerel WHERE agent_id = 7 INTERSECT ALL SELECT agent_id FROM skillrel WHERE agent_id = 7 agent_id 7 7 Agent 7 has at most 2 of either skills or languages or both CS386/586 Introduction to Database Systems, Lois Delcambre Slide 2

3 Except vs. Except all SELECT agent_id FROM languagerel WHERE agent_id = 7 EXCEPT SELECT agent_id FROM skillrel WHERE agent_id = 7 No rows found. Agent 7 has both languages and skills SELECT agent_id FROM languagerel WHERE agent_id = 7 EXCEPT ALL SELECT agent_id FROM skillrel WHERE agent_id = 7 agent_id 7 7 Agent 7 has two more languages than skills CS386/586 Introduction to Database Systems, Lois Delcambre Slide 3

4 Except vs. Except all (cont.) SELECT agent_id FROM languagerel WHERE agent_id = 7 EXCEPT ALL SELECT agent_id FROM skillrel WHERE agent_id = 7 agent_id (repeated from previous slide) 7 Agent 7 has two more languages 7 than skills SELECT agent_id FROM skillrel WHERE agent_id = 7 EXCEPT ALL SELECT agent_id FROM languagerel WHERE agent_id = 7 No rows found. Agent 7 has fewer skills than languages. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 4

5 Union vs. Union ALL SELECT agent_id FROM languagerel WHERE agent_id = 7 UNION SELECT agent_id FROM skillrel WHERE agent_id = 7 agent_id Agent 7 has languages or skills 7 SELECT agent_id FROM languagerel WHERE agent_id = 7 UNION ALL SELECT agent_id FROM skillrel WHERE agent_id = 7 Agent 7 s languages + skills = 6 agent_id CS386/586 Introduction to Database Systems, Lois Delcambre Slide 5

6 Relational Algebra (as defined mathematically by Codd) vs. SQL σ, Π,, X, ρ for relations as sets σ, Π,, X, ρ for tables as bags,, for relations as sets,, for tables as bags Relational Algebra Yes NO Yes NO SQL GROUP BY/HAVING NO YES YES (with DISTINCT) Basic Select-From-Where YES (it s the default) Basic Select-From-Where YES (UNION, INTERSECT, EXCEPT) YES (UNION ALL, INTERSECT ALL, EXCEPT ALL) Aggregate operators NO YES (divide operator) YES NO (not directly) Note that in a DBMS, there are algebraic operators for bags/sets and for grouping and aggregates. We use (only) the original rel. alg. operators defined for sets. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 6

7 How do you like to see the Spy Database? As tables? Agent(agent_id,first,middle,last,address,city,country,salary,clearance_id) Mission(mission_id,, access_id, team_id, mission_status) Affiliation(aff_id, title, description) AffiliationRel(aff_id, agent_id, affiliation_strength) LanguageRel(lang_id, agent_id) Language(lang_id, language) SecurityClearance(sc_id, sc_level, description) Skill(skill_id, skill) SkillRel(skill_id, agent_id) Team(team_id,, meeting_frequency) TeamRel(team_id, agent_id) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 7

8 How do you like to see the Spy Database? As tables with foreign keys? Agent(agent_id,first,middle,last,address,city,country,salary,clearance_id) Mission(mission_id,, access_id, team_id, mission_status) Affiliation(aff_id, title, description) AffiliationRel(aff_id, agent_id, affiliation_strength) LanguageRel(lang_id, agent_id) Language(lang_id, language) SecurityClearance(sc_id, sc_level, description) Skill(skill_id, skill) SkillRel(skill_id, agent_id) Team(team_id,, meeting_frequency) TeamRel(team_id, agent_id) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 8

9 How do you like to see the Spy Database? MS Access Relationship Diagram of tables? Affiliation aff_id AffiliationRel aff_id agent_id affiliation_strength SkillRel skill_id agent_id Skill skill_id Language lang_id LanguageRel lang_id agent_id Agent agent_id clearance_id TeamRel team_id agent_id SecurityClearance clearance_id Team team_id Mission mission_id team_id access_id CS386/586 Introduction to Database Systems, Lois Delcambre Slide 9

10 How do you like to see the Spy Database? Entity-Relationship Diagram (ERD)? Affiliation strength AffiliationRel SkillRel TeamRel Language Agent Team 0..1 LanguageRel Skill 1..1 SecurityClearance 1..1 Mission Rectangles: Entities Lines: Relationships (with cardinalities) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 10

11 Entity-Relationship Diagram (ERD) Note: Relationships need s. (I ll use the ones from the Spy DB tables for the manymany relationship.) Language Skill LanguageRel AffiliationRel SkillRel Affiliation Agent SecurityClearance Team strength TeamRel Agent-clearance 1..1 Other relationships arise from foreign keys; they needs s, too. Mission Mission-clearance AssignedTeam CS386/586 Introduction to Database Systems, Lois Delcambre Slide 11

12 Conceptual Database Design Using the Entity-Relationship (ER) Model CS386/586 Introduction to Database Systems, Lois Delcambre Slide 12

13 Overview of Database Design Conceptual design: (ER Model is used for this.) What are the entities and relationships we need? Logical design: Transform ER design to Relational Schema Schema Refinement: (Normalization) Check relational schema for redundancies and related anomalies. Physical Database Design and Tuning: Consider typical workloads; (sometimes) modify the database design; select file types and indexes. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 13

14 Entity-Relationship Model is a different model than the Relational Model Relation model has: tables(relations) with attributes, keys, foreign keys, domain definitions for attributes Entity-Relationship model has: Entities and entity sets with attributes, keys, and domain definitions for attributes relationships among entities and relationship sets with cardinality constraints (in the book they refer to key constraints and participation constraints) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 14

15 Entity-Relationship Diagram (original syntax) code ssn lot home Department Legend: Entity set Relationship set Attribute manager sponsor assigned Project number start date end date budget CS386/586 Introduction to Database Systems, Lois Delcambre Slide 15

16 Definitions Entity: Real-world object distinguishable from other objects. (An employee d John Smith, for example.) An entity is described using a set of attributes. Entity Set: A collection of similar entities. (All employees in the company, for example.) (An entity set if often referred to as just an entity if everyone knows we re talking about the schema rather than individual rows. An entity set may also be called entity type.) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 16

17 Definitions Relationship: Association among 2 or more entities. John s home department is Pharmacy, for example. John is an entity. Pharmacy is an entity. Relationship Set: Collection of similar relationships. The home department relationship set consists of all indications of a home department for an employee. (A relationship set is often referred to as a relationship because everyone knows we are talking about the schema rather than individual rows. A relationship set may also be called a relationship type.) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 17

18 UML version of the same E-R Diagram UML: Unified Modeling Language for OO Design ssn lot home manager Department code assigned Project number start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 18

19 Equivalent Relational Schema extra table to represent a many-to-many relationship (ssn,, lot, home-dept) Project-team(ssn, number) Department (id,, manager) Project (number,, start-date, end-date, budget, sponsor) ssn lot assigned home manager Project number start-date end-date budget Department code sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 19

20 Equivalent Relational Schema - with foreign keys shown (ssn,, lot, home-dept) Project-team(ssn, number) Department (id,, manager) Notice that the many-to-many relationship set must be represented in a (new) table. Project (number,, start-date, end-date, budget, sponsor) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 20

21 ERD Relational DB Diagram Many-to-many relationship sets ssn lot ssn lot assigned assigned-projects Proj-Team SSN PNumber team home manager Project number start-date end-date budget home manager Project number start-date end-date budget Department code sponsor Department code sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 21

22 What data do we need to record a relationship? ssn lot home Department code We must indicate which employee and which department are to be connected (for each relationship). We need the key value for an employee and the key value for the department stored together to represent the relationship. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 22

23 Cardinality Constraints on Relationship sets: How many entities can participate? ssn lot home manager Department code assigned Project number start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 23

24 Cardinality Constraints on Relationship sets How many entities can participate? ssn lot 1..1 home manager Department code 1..1 assigned Project number start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 24

25 An employee can have 0 or 1 home departments ssn lot 1..1 home manager Department code 1..1 assigned Project number start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 25

26 A department can have 0 to any number of employees ssn lot 1..1 home manager Department code 1..1 assigned Project number start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 26

27 Unified Modeling Language (UML): Class Diagram ssn lot 1..1 home Departments code ssn lot home 0..1 Departments code ssn lot home 1..* Departments code Which one is right? We must discover the semantics of the application! CS386/586 Introduction to Database Systems, Lois Delcambre Slide 27

28 Various notation for one-to-many one many 1 m 1 * zero..one one..many * maximum cardinalities only minimum and maximum cardinalities Notice, UML notation allows any sort of cardinality. DBMSs are limited to enforcing a max of 1 or many and a min of 0 or 1. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 28

29 Various notations for many-to-many many many one..many one..many 1..* 1..* m n * * maximum cardinalities only minimum and maximum cardinalities CS386/586 Introduction to Database Systems, Lois Delcambre Slide 29

30 Relationship sets can have attributes ssn lot start-date descriptive attribute of the relationship set code E-R notation home Department UML notation ssn lot home 0..1 Department code start-date CS386/586 Introduction to Database Systems, Lois Delcambre Slide 30

31 Try all three locations for the attributes: which one makes sense? ssn lot home 0..1 start-date Department code? start-date? start-date? CS386/586 Introduction to Database Systems, Lois Delcambre Slide 31

32 Relationship sets can have role s (in addition to the of the relationship set) ssn lot manager managed-by manages Department code role role relationship set CS386/586 Introduction to Database Systems, Lois Delcambre Slide 32

33 Example: reading role s An employee manages 0 or 1 departments ssn lot manager managed-by manages Department code role role relationship set CS386/586 Introduction to Database Systems, Lois Delcambre Slide 33

34 Same entity sets can participate in different roles for the same relationship set E-R notation UML notation ssn supervisor lot subordinate ssn lot 0..1 supervisor subordinate Reports-to Reports_To CS386/586 Introduction to Database Systems, Lois Delcambre Slide 34

35 Exercise: Draw an ERD Professors have a SSN, a and an age and their SSNs uniquely identify them. Professors teach courses. Each course is supervised by one professor. Courses are uniquely identified by their courseid, and they have a. Choose cardinalities for all your relationships. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 35

36 Weak Entity Sets (and Identifying Relationship sets) Department code 0..1 home manager 1 SSN office strong entity set This is a weak entity set It wouldn t be in the database if the strong entity weren t present. children Dependents date-of-birth identifying relationship set weak entity set CS386/586 Introduction to Database Systems, Lois Delcambre Slide 36

37 Weak Entity Sets and Identifying Relationship sets: Alternative Notation Department code * home * manager 1 SSN office strong entity set participation in the identifying relationship set is required partial key - must be concatenated with key for the strong entity set (SSN) Children Dependents date-of-birth identifying relationship set weak entity set CS386/586 Introduction to Database Systems, Lois Delcambre Slide 37

38 Weak Entity Sets and Identifying Relationship sets: Alternative Notation (cont.) Department code * home * manager 1 Cardinalities for an identifying relationship set SSN office 1..1 Children Dependents Name Date-of-birth strong entity set identifying relationship set weak entity set CS386/586 Introduction to Database Systems, Lois Delcambre Slide 38

39 Ternary vs. Binary Relationship sets Product Supplier Product supplies Supplier supplies required-for authorized -for Project Project These two schemas are not equivalent! When would we use a ternary relationship set? When would we use three binary relationship sets? CS386/586 Introduction to Database Systems, Lois Delcambre Slide 39

40 Binary vs. Ternary Relationship sets (Cont.) The ternary relationship set means that a Supplier must be authorized to supply a particular part to a particular project. e.g., Office-Depot can supply laser printer paper to project 112. Office-Max can supply paper clips to Project 112. Office-Max can supply pencils to project 115. (But based on that much information, Office-Max can t supply pencils to 112.) The three binary relationship sets each represent something distinct. A Supplier can be authorized to supply certain products (Office-Max can supply pencils). A Project can require certain products (112 needs pencils). And a Supplier can be authorized to supply a certain project. (Office-Max supplies 112) Therefore: Office-Max can supply pencils to 112. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 40

41 Duality: entity value and attribute relationship Project P-number P- Due-Date Assignment Manager SSN E-Name Office Should Office be an attribute of? or a separate entity set? Most attributes can be promoted to an entity set and some entities can be demoted to an attribute value. This explains why there are so many different ways to design a schema. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 41

42 What are some reasons to model Office as an entity set? Project P-number P- Due-Date Assignment Manager SSN E-Name Assigned Office Office O-number an employee can have more than one office there are other attributes of Office Office needs to participate in other relationship sets such as a relationship set connecting to furniture or telephones or network drops (located in the office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 42

43 Office as an entity (with attributes & relationships) Project P-number P- Due-Date Assignment Manager SSN E-Name Assigned Office Office O-number sq-ft orientation Location NetworkDrop IP-address location CS386/586 Introduction to Database Systems, Lois Delcambre Slide 43

44 What are some reasons to model Office as an attribute of? Project P-number P- Due-Date Assignment Manager SSN E-Name O-number Assigned Office Office O-number it is faster to access office number; there is no join required. schema is a little simpler; one less table CS386/586 Introduction to Database Systems, Lois Delcambre Slide 44

45 Key Constraints - as described in the book (limit participation in relationship set to at most 1) same as maximum multiplicity of 1 in UML d- ssn lot d-code Manages Department ssn lot 1 Manages * Department d-code d Each dept has at most one manager, according to the key constraint on Manages. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 45

46 Participation Constraint - as in text: when every entity MUST participate in a relationship set a Project has exactly one manager Project Manager 1..1 Project n 1 Manager a Project MUST have a manager and there is at most 1 employee who is manager CS386/586 Introduction to Database Systems, Lois Delcambre Slide 46

47 Translating an ER Diagram to a Relational Schema CS386/586 Introduction to Database Systems, Lois Delcambre Slide 47

48 1. Translate each entity set into a table, with keys. Entity set: can be represented as a table in the relational model SSN E-Name Office has a key which becomes a key for the table CREATE TABLE (SSN CHAR(11), E-Name CHAR(20), Office INTEGER, PRIMARY KEY (SSN)) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 48

49 If your relational DBMS does not allow multi-valued attributes then 1a. Create a table for each multi-valued attribute. How many offices can one employee have? Project (P-number, P-, Due-Date) (SSN, E-Name, Office) vs. Project (P-number, P-, Due-Date) (SSN, E-Name) Office-Assignment (SSN, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 49

50 Sample Data Project (P-number, P-, Due-Date) (SSN, E-Name, Office) 12 Smith O Wei O Jones O-112 Project (P-number, P-, Due-Date) (SSN, E-Name) 12 Smith 15 Wei Office-Assignment (SSN, Office) 12 O O O-110 CS386/586 Introduction to Database Systems, Lois Delcambre Slide 50

51 2. Translate each many to many relationship set into a table Project P-number P- Due-Date * Assignment * Manager 1 SSN E-Name Office What are the attributes and what is the key for Assignment? Assignment (?) Project (P-number, P-, Due-Date) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 51

52 Project P-number P- Due-Date * Assignment * Manager 1 SSN E-Name Office Answer: Assignment (P-Number, SSN) P-Number is a foreign key for Project SSN is a foreign key for Project (P-Number, P-Due-Date) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 52

53 What should we do with each one to many relationship set? Project P-number P- Due-Date * Assignment * Manager 0..1 SSN E-Name Office What should we do with each one to many relationship set? Manager (?) Project (P-number, P-, Due-Date) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 53

54 Answer: Create a foreign key for a 1-tomany relationship set (on the many side). Project P-number P- Due-Date * Assignment * Manager 1 SSN E-Name Office Project (P-number, P-, Due-Date, Manager) (SSN, E-Name, Office) Manager is a foreign key (referencing the relation) value of Manager must match an SSN CS386/586 Introduction to Database Systems, Lois Delcambre Slide 54

55 Or... create a table for a 1-many relationship set. Project P-number P- Due-Date * Assignment * Manager 1 SSN E-Name Office Project (P-number, P-, Due-Date, Manager) (SSN, E-Name, Office) vs. Project (P-number, P-, Due-Date) (SSN, E-Name, Office) Manager (P-number, SSN) What are the tradeoffs between these two? Note: P-number is the key for Manager CS386/586 Introduction to Database Systems, Lois Delcambre Slide 55

56 Project P-number P- Due-Date * Assignment * Manager 1 SSN E-Name Office Project (P-number, P-, Due-Date) (SSN, E-Name, Office, Managed-project) vs. Project (P-number, P-, Due-Date) (SSN, E-Name, Office) Manager (P-number, SSN) What if SSN is the key for Manager? CS386/586 Introduction to Database Systems, Lois Delcambre Slide 56

57 What do we do when a many-to-many relationship set has an attribute? Project P-number P- Due-Date Assignment role start-date end-date SSN E-Name Office Assignment (A-project, A-SSN) Project (P-number, P-, Due-Date) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 57

58 Add attributes to the table representing the relationship Project P-number P- Due-Date Assignment role start-date end-date SSN E-Name Office Assignment (A-project, A-SSN, role, start-date, end-date) Project (P-number, P-, Due-Date) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 58

59 What do we do when a 1-to-many relationship set has an attribute? Project P-number P- Due-Date Manager start-date end-date SSN E-Name Office Project (P-number, P-, Due-Date, Manager) (SSN, E-Name, Office) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 59

60 Add them to the table with the foreign key for that represents the relationship Project P-number P- Due-Date Manager start-date end-date SSN E-Name Office What do we do when a 1-to-many relationship set has an attribute? Project (P-number, P-, Due-Date, Manager, start-date, end-date) (SSN, E-Name, Office ) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 60

61 Participation Constraints in SQL We can require any table to be in a binary relationship using a foreign key which is required to be NOT NULL (but little else without resorting to CHECK constraints) CREATE TABLE Department ( d-code INTEGER, d- CHAR(20), manager-ssn CHAR(9) NOT NULL, since DATE, PRIMARY KEY (d-code), FOREIGN KEY (manager-ssn) REFERENCES, ON DELETE NO ACTION) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 61

62 Translating Weak Entity Sets Weak entity sets and identifying relationship sets are translated into a single table. Must include key of strong entity set, as a foreign key. When the owner entity is deleted, all owned weak entities must also be deleted. CREATE TABLE Insurance_Policy ( dep- CHAR(20), age INTEGER, cost REAL, ssn CHAR(11) NOT NULL, PRIMARY KEY (dep-, ssn), FOREIGN KEY (ssn) REFERENCES, ON DELETE CASCADE) CS386/586 Introduction to Database Systems, Lois Delcambre Slide 62

63 Note ERDs and UML Diagrams can be at two levels: the ERD/UML level and the Relational Table level. The difference is primarily with the many-to-many relationship sets. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 63

64 Summary of Translation Steps: ER to Tables 1. Create table and choose key for each entity set; include single-valued attributes. 2. Create table for each weak entity set; include single-valued attributes. Include key of owner as a foreign key in the weak entity. Set key as foreign key of owner plus local, partial key. 3. For each 1:1 relationship set, add a foreign key to one of the entity sets involved in the relationship (a foreign key to the other entity in the relationship)*. 4. For each 1:N relationship set, add a foreign key to the entity set on the N-side of the relationship (to reference the entity set on the 1-side of the relationship)*. 5. For each M:N relationship set, create a new table. Include a foreign key for each participant entity set, in the relationship set. The key for the new table is the set of all such foreign keys. 6. For each multi-valued attribute, construct a separate table. Repeat the key for the entity in this new table. It will serve as both the key for this table as well as a foreign key to the original table for the entity. * Unless relationship set has attributes. If it does, create a new table for the relationship set. This algorithm from Elmasri/Navathe, p CS386/586 Introduction to Database Systems, Lois Delcambre Slide 64

65 Entity-Relationship Diagram ssn lot 1..1 home manager Department d-code d assigned Project p-number p- start-date end-date budget sponsor CS386/586 Introduction to Database Systems, Lois Delcambre Slide 65

66 Equivalent Relational Schema ssn lot home-dept 1 1 home manager 1 Department d-code d- manager 1 Assignment SSN p-number 1 Project p-number p- start-date end-date budget sponsoring-dept sponsor Notice that the relationship sets shown in this diagram aren t really needed. foreign keys reference other tables. CS386/586 Introduction to Database Systems, Lois Delcambre Slide 66