Download presentation
Presentation is loading. Please wait.
Published byAmie Louisa Todd Modified over 9 years ago
1
ITEC 3220M Using and Designing Database Systems Instructor: Prof. Z.Yang Course Website: http://people.math.yorku.ca/~zyang/itec 3220m.htm http://people.math.yorku.ca/~zyang/itec 3220m.htm Office: Tel 3049
2
Chapter 3 The Relational Database Model (Cont’d)
3
3 Relational Database Operators Relational algebra –Defines theoretical way of manipulating table contents using relational operators: SELECT PROJECT JOIN INTERSECT –Use of relational algebra operators on existing tables (relations) produces new relations UNION DIFFERENCE PRODUCT DIVIDE
4
4 Relational Algebra Operators ( continued ) Union: –Combines all rows from two tables, excluding duplicate rows –Tables must have the same attribute characteristics Intersect: –Yields only the rows that appear in both tables
5
5 Union
6
6 Intersect
7
7 Relational Algebra Operators ( continued ) Difference –Yields all rows in one table not found in the other table—that is, it subtracts one table from the other
8
8 Venn Diagrams for Traditional Set Operators UnionIntersection Venn Diagram: Differences
9
9 Product Yields all possible pairs of rows from two tables
10
10 Relational Algebra Operators ( continued ) Select –Yields values for all rows found in a table –Can be used to list either all row values or it can yield only those row values that match a specified criterion –Yields a horizontal subset of a table Project –Yields all values for selected attributes –Yields a vertical subset of a table
11
11 Select
12
12 Project
13
13 Relational Algebra Operators ( continued ) Join –Allows us to combine information from two or more tables –Real power behind the relational database, allowing the use of independent tables linked by common attributes
14
14 Links tables by selecting rows with common values in common attribute(s) Three-stage process –Product creates one table –Select yields appropriate rows –Project yields single copy of each attribute to eliminate duplicate columns Natural Join Process
15
15 Natural Join ( continued ) Final outcome yields table that –Does not include unmatched pairs –Provides only copies of matches If no match is made between the table rows, –the new table does not include the unmatched row
16
16 Other Joins EquiJOIN –Links tables based on equality condition that compares specified columns of tables –Join criteria must be explicitly defined Theta JOIN –EquiJOIN that compares specified columns of each table using operator other than equality one Outer JOIN –Matched pairs are retained –Unmatched values in other tables left null –Right and left
17
17 Divide Requires use of single-column table and two-column table
18
18 Summary of Meanings of the Relational Algebra Operators Select: Extracts rows that satisfy a specified condition Project: Extracts specified columns Product: Builds a table from two tables consisting of all possible combinations of rows, one from each of the two tables Union: Builds a table from all rows appearing in either of two tables Intersect: Builds a table consisting of all rows appearing in both of two specified tables
19
19 Summary of Meanings of the Relational Algebra Operators (Cont’d) Join: Extracts rows from a product of two tables such that two input rows contributing to any output row satisfy some specified condition Outer Join: Extracts the matching rows of two tables and the unmatched rows from both tables Divide: Builds a table consisting of all values of one column of a binary table that match all values in a unary table
20
Chapter 4 Entity Relationship (E-R) Modeling
21
21 In this chapter, you will learn: How relationships between entities are defined and refined, and how such relationships are incorporated into the database design process Key terms: cardinality, connectivity, optional, mandatory, strong relationship, weak relationship, supertype, subtype, etc. How to develop an E-R diagram
22
22 The Entity Relationship (ER) Model ER model forms the basis of an ER diagram ERD represents the conceptual database as viewed by end user ERDs depict the ER model’s three main components: –Entities –Attributes –Relationships
23
23 Entities Refers to the entity set and not to a single entity occurrence Corresponds to a table and not to a row in the relational environment In both the Chen and Crow’s Foot models, an entity is represented by a rectangle containing the entity’s name Entity name, a noun, is usually written in capital letters
24
24 Attributes Characteristics of entities Domain is set of possible values Primary keys underlined
25
25 Examples EMPLOYEE (EMPLOYEE _ID, EMPLOYEE _NAME, ADDRESS, DATE-EMPLOYED) ADDRESS EMPLOYEE _NAME EMPLOYEE EMPLOYEE _ID DATE- EMPLOYED EMPLOYEE EMPLOYEE _ID EMPLOYEE _NAME ADDRESS DATE- EMPLOYED
26
26 Attributes (Cont’d) Simple –Cannot be subdivided –Age, sex, marital status Composite –Can be subdivided into additional attributes –Address into street, city, zip Single-valued –Can have only a single value –Person has one social security number Multi-valued –Can have many values –Person may have several college degrees –In the Chen E-R model, the multivalued attributes are shown by a double line connecting the attributes to the entity Derived –Can be derived with algorithm –Age can be derived from date of birth –Versus stored attribute
27
27 Attributes (Cont’d) An attribute broken into component parts Address Street_Address Post_Code StateCity
28
28 Attributes (Cont’d) Years_Employed Employee_ID Date_ Employed Skills AddressEmployee _Name EMPLOYEE Entity with a multivalued attribute (Skill) and derived attribute (Years_Employed)
29
29 How to Deal with Multivalued Attributes With the original entity, create several new attributes, one for each of the original multivalued attribute’s components. Create a new entity composed of the original multivalued attribute’s components.
30
30 An Example Car_Vin Mod_code Car_Color Car_Year CAR
31
31 Relationships Association between entities Connected entities are called participants Operate in both directions Connectivity describes relationship classification –1:1, 1:M, M:N Cardinality –Expresses number of entity occurrences associated with one occurrence of related entity
32
32 ERD Symbols Rectangles represent entities Diamonds represent the relationship(s) between the entities “1” side of relationship –Number 1 in Chen Model –Bar crossing line in Crow’s Feet Model “ Many” relationships –Letter “M” and “N” in Chen Model –Three pronged “Crow’s foot” in Crow’s Feet Model
33
33 Connectivity and Cardinality in an ERD
34
34 Relationship Strength Existence dependence –Entity’s existence depends on existence of related entities –Existence-independent entities can exist apart from related entities –EMPLOYEE claims DEPENDENT Weak (non-identifying) –One entity is existence-independent on another –PK of related entity doesn’t contain PK component of parent entity Strong (identifying) –One entity is existence-dependent on another –PK of related entity contains PK component of parent entity
35
35 Weak Entity Existence-dependent on another entity Has primary key that is partially or totally derived from parent entity
36
36 Relationship Participation Optional –Entity occurrence does not require a corresponding occurrence in related entity –Shown by drawing a small circle on side of optional entity on ERD Mandatory –Entity occurrence requires corresponding occurrence in related entity –If no optionality symbol is shown on ERD, it is mandatory
37
37 Relationship Degree Indicates number of associated entities Unary –Single entity –Exists between occurrences of same entity set Binary –Two entities associated –Most common –To simplify the conceptual design, most higher-order relationships are decomposed into appropriate equivalent relationships when possible Ternary –Three entities associated
38
38 Three Types of Relationships
39
39 Recursive Relationship Definition: A relationship can exist between occurrences of the same entity set. PERSON is married to 11 EMPLOYEE manages 1M
40
40 Composite Entities Also known as bridge entities Composed of the primary keys of each of the entities to be connected May also contain additional attributes that play no role in the connective process
41
41 A Composite Entity in an ERD
42
42 Example M:N Relationship
43
43 Converting M:N Relationship to Two 1:M Relationships (Cont’d)
44
44 An Example claims DEPENDENT ORDER PRODUCT EMPLOYEE employs STORE 1ZXYW 1MM (a,b)(g,h)(i,j) (c,d) (e,f)(k,l) (o,p) (m,n)
45
45 Comparison of E-R Modeling Symbols
46
46 Developing an E-R Diagram Iterative Process –Step1: General narrative of organizational operations developed –Step2: Basic E-R Model graphically depicted and reviewed –Step3: Modifications made to incorporate newly discovered E-R components Repeat process until designers and users agree E-R Diagram complete
47
47 Example Create an ERD using the following business rules: –A company operates four departments –Each department employs employees –Each of the employees may or may not have one or more dependents –Each employee may or may not have an employment history
48
48 Exercise Design an E-R diagram for a real estate firm that lists property of sale. l The firm has a number of sales offices in several states. l Each sales office is assigned one or more employees. Attributes of employees include ID and name. An employee must be assigned to only one sales office. l For each sales office, there is always one employee assigned to manage that office. An employee may manage only the sales office to which he is assigned. l The firm lists property for sale. Attributes of property include ID and location. Components of location include address, city, state, and Zip_code. l Each unit of property must be listed with one of the sales offices. A sales office may have any number of properties listed, or may have no properties listed. l Each unit of property has one or more owners. An owner may own one or more units of property. An attribute of the relationship between property and owner is Percent_Owned.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.