COMP 7990

Principles and Practices of

Data Analytics

Data Management - I

Dr. Kevin Wang

Outline

▪ Introduction to Databases

▪ Relational Data Model

▪ Introduction to SQL
Why Database?

▪ The purpose of a database is:

– To store data
– To provide an organizational structure for data
– To provide a mechanism for interacting with data
➢ Create
➢ Read
➢ Update
➢ Delete

▪ A database can store information and relationships that are more

complicated than storing data in a simple file.
Problems for storing data in a simple file: Redundancy

▪ Problems for storing data in a simple file

– In a file, each row stands for one record for its own. As a results, the same information
may be entered serval times.
➢ For example, we use a file to track of the instructor of each course. This file may include the
course instructor’s name, ID, office.
➢ If an instructor is currently teaching 4 courses, his/her information would appear in the file 4
times.
➢ It costs more storage space than necessary.
Problems with storing data in a simple file: Multiple
Business Concepts
▪ Each row in a file may contain information on more than one business
concept.

▪ For example, a file of courses information may include instructor information

(Instructor Name, InstructionID, office) and course information (CourseID,
CourseName) in the same row.
File Modification Issues
多余的
▪ Redundancy and multiple business concepts create modification problems
– Deletion Problems (Deletion Anomalies)
异常

Suppose the course COMP4405 is canceled.

Deleted this row → Instructor’s information is also lost
File Modification Issues

▪ Redundancy and multiple business concepts create modification problems

– Deletion Problems (Deletion Anomalies)
– Update Problems (Update Anomalies) If instructor Zhang, Lu is changed to LIU, YANG
for course COMP4115, we need to also change
Instructor ID, Office and Phone Number as well.
File Modification Issues

▪ Redundancy and multiple concepts create modification problems

– Deletion Problems (Deletion Anomalies) Insert a row for information about a new
– Update Problems (Update Anomalies) instructor → course ID and CourseName
information is missing.
– Insertion Problems (Insertion Anomalies)
Relational Databases

▪ Relational databases are designed to address many of the information

complexity issues that arise in business.

▪ A relational database stores information in tables. Each concept is stored in

its own table.

▪ In essence, a relational database will break up a file into several parts

(tables).
– One part for each concept in the file.
– For example, a file of courses might be divided into a INSTRUCTOR table, a COURSE
table and a COURSE_INSTRUCTOR table.
Relational Databases (con't)

▪ There is another type of data storage model used in applications like big data
and machine learning.

▪ NoSQL databases store data in the same big table.

▪ NoSQL is useful when

• Data structure is simple
• The amount of data is extremely large (efficiency is important)
 COURSE COURSE_INSTRUCTOR INSTRUCTOR
Course ID CourseName Course ID Instructor ID Instructor Instructor ID Office
COMP7990 Principles and Practices of Data COMP7990 3 Zhang, Lu 3 RRS708
Analytics
COMP4125 3
COMP4125 Visual Analytics LIU, Yang 2 RRS631
COMP7820 3
COMP7820 Visual Analytics and Decision COMP4115 3
Support
COMP4405 2
COMP4115 Exploratory Data Analysis and
Visualization
COMP4405 Data Mining
 Relational Databases

▪ The information for courses and instructors can be joined back together.
▪ In a relational database, tables are joined together using matched pairs of
data values.
– For example, the InstructorID is stored in a column in COURSE_INSTRUCTOR table.
Whenever we need information about an instructor, we can use InstructorID to look up
the instructor information in the INSTRUCTOR Table.
COURSE COURSE_INSTRUCTOR INSTRUCTOR

Course ID CourseName Course ID Instructor ID Instructor Instructor ID Office

COMP7990 Principles and Practices of Data COMP7990 3 Zhang, Lu 3 RRS708
Analytics
COMP4125 3
COMP4125 Visual Analytics LIU, Yang 2 RRS631
COMP7820 3
COMP7820 Visual Analytics and Decision COMP4115 3
Support
COMP4405 2
COMP4115 Exploratory Data Analysis and
Visualization
COMP4405 Data Mining
Course ID CourseName Instructor Instructor Office Phone
ID Number
COMP7990 Principles and Practices of
Data Analytics LAN, Liang 3 RRS708 34115880
COMP4125 Visual Analytics LIU, Yang 2 RRS631 34112798
COMP7820 Visual Analytics and INSTRUCTOR
Decision Support
COMP4115 Exploratory Data Analysis
and Visualization
COMP4405 Data Mining
COURSE

Course ID Instructor ID
COMP7990 3
COMP4125 3
COMP7820 3
COMP4115 3
COMP4405 2
COURSE_INSTRUCTOR
Relational Databases

▪ You may think databases make things complicated.

– Breaking file into tables
– Obtain more complete information by joining tables (maybe some information is useless)
冗余
▪ However, a relational database minimizes data redundancy, preserves
complex relationships among concepts, and allows for partial data (null values).

▪ Furthermore, a relational database provides a solid foundation to create

user interface forms and reports.
– Structured Query Language (SQL)
➢ An international standard language for creating, processing and querying databases and their
tables.
➢ A vast majority of data-driven applications and websites use SQL.
Database System

▪ The four components of a database system

– Users

– Database application(s)

– Database Management System (DBMS)

– Databases
Database System

▪ Users
▪ Database application(s)
– User interacted with database
applications (e.g. website)
▪ Database Management System
(DBMS)
– A gate-keeper for database. The
database application needs go through
DBMS to interact with the database
▪ Databases
– Store raw data in separate tables
User

▪ A user of a database system will:

– Use a database application to keep track of information
(e.g, checking bank balance)

– Use different user interface forms to enter, read, delete and query data

– Produce reports
Database Applications

▪ A database application is a set of one or more computer programs or

websites that serve as an intermediary between the user and the DBMS.
中间人
– Data-driven website
– Mobile apps

▪ These application can not directly access data in database. They need to go
through DBMS to touch the data.
Database Management System (DBMS) Oracle

▪ A database management system (DBMS) serves as an intermediary

(gatekeeper) between database applications and the database.

▪ The DBMS manages and controls database activities.

▪ The DBMS creates, processes and administers the databases it controls

 The Database

▪ A database is a self-describing Integer

collection of related records

▪ Self-describing
Instructor Instructor Office Phone
– The database itself contains the definition of ID Number
its structure.
元数据 – Metadata are data describing the structure of Zhang, Lu. 3 RRS708 34115880
data in the database (e.g, the data type for LIU, Yang 2 RRS631 34112798
Instructor ID is Integer)

▪ Tables with a relational database are

related to each other in some way
Database Contents

• User data
• Metadata
• Indexes and other overhead data
• Application metadata
Functions of a DBMS

▪ Create databases
▪ Create tables
▪ Create relationships between tables
▪ Read database data
▪ Modify database data (insert, update, delete)
▪ Maintain database structures
▪ Enforce rules
▪ Control concurrency 并发控制
▪ Provide security
▪ Perform data backup and recovery
 Enforce Rules: Referential Integrity Constraints

▪ A DBMS can enforce many type of constraints.

▪ One of the most useful type is Referential Integrity Constraints. 参照完整性约束
▪ Referential integrity constraints ensure that the values of a column in one table
are valid based on the values in another table.
– For example, if 4 was entered as a InstructorID in the COURSE_INSTRUCTOR table, a Instructor having a
InstructorID value of 4 MUST exist in INSTRUCTOR table.

– If Instructor ID 4 does not exist, the DBMS would not allow us to add the record.
Course ID Instructor ID
Instructor Instructor ID Office
COMP7990 3
Zhang, Lu 3 RRS708
COMP4125 3
LIU, Yang 2 RRS631
COMP7820 3
New record COMP4115 3
COMP4405 2 Does InstructorID 4 exist in INSTRUCTOR table?
COMP4005 4
Outline

▪ Introduction to Databases

▪ Relational Data Model

▪ Introduction to SQL
Entity 实体、对象

▪ An entity is something of importance to a user of organization that needs to

be represented in a database.

▪ Database entity is a thing, person, place, unit, object or any item about which the
data should be captured and stored in the form of properties, workflow and tables.

▪ An entity represents one concept.

–e.g., course, instructor …

Instructor ID Instructor Office

3 Zhang, Lu. RRS708

2 LIU, Yang RRS631

Relation

▪ A relation is a two-dimensional table that has specific characteristics

– Rows contain data about instances of an entity
– Columns contain data about attributes of the entity
– Cells of the table hold a single value
– All values in a column are of the same data type
– Each column has a unique name
– The order of the columns is unimportant Each column is an attribute.
– The order of the rows is unimportant
– No two rows can be identical
完全相同的

InstructorID InstructorName Office Phone Number

Each row represents an instructor 3 Zhang, Lu RRS708 34115880

2 LIU, Yang RRS631 34112798

A Non-relation Example

InstructorID InstructorName Office Phone Number

3 Zhang, Lu RRS708 34115880

2 LIU, Yang RRS631 34112798

4 ZHANG, Wei RRS707 34112001,51091111

3 Zhang, Lu RRS708 34115880

 A Non-relation Example

InstructorID InstructorName Office Phone Number

3 Zhang, Lu RRS708 34115880

2 LIU, Yang RRS631 34112798

Identical rows 4 ZHANG, Wei RRS707 34112001,51091111 Multiple values in a cell

3 Zhang, Lu RRS708 34115880

All relations are tables. However, not all tables are relations
Synonyms 同义词

Table Row Column

Relation Record Field

File (rarely) Tuple (rarely) Attribute

Keys

▪ A key is one (or more) columns of a relation whose values are used to
identify a row

Unique Key Non-unique Key

Data value is unique for each row. Data value may be shared among
multiple rows.
Therefore, the key will uniquely
identify a row. Therefore, the key will identify a set of
rows.
Example

InstructorID InstructorName Office Phone Department ID

Number
3 Zhang, Lu RRS708 34115880 101

2 LIU, Yang RRS631 34112798 101

9 WANG, Alan AAB556 34110001 102

Unique Key Non-unique Key

 Keys

Unique Keys Non-Unique Keys

Candidate Composite Primary Surrogate Foreign

Key Key Key Key Key
Keys

▪ Unique Keys
– Candidate Key
➢ A candidate key is called “candidate” because it has the potential to become the primary key.
– Primary Key
➢ A primary key is a candidate key chosen to be the main key for the relation.
– Composite Key 复合键
唯一性
➢ A composite key is a key that is composed of two or more attributes for uniqueness.
➢ e.g., Flight number + Date
– Surrogate Key 代理键
➢ A surrogate key is a unique, numeric value that is added to a relation to serve as the primary
key.
➢ Surrogate key values have no meaning to users.
➢ e.g., Instructor ID
▪ Non-unique Keys
– Foreign Key
Primary Key

▪ If you know the value of the primary key, you will be able to uniquely identify
a single row within table.

InstructorID InstructorName Office Phone

Number

3 Zhang, Lu RRS709 34115880

2 LIU, Yang RRS631 34112798
9 WANG, Alan AAB556 34110001

Primary Key
Composite Key

▪ A composite key is a key that combines two or more attributes for

uniqueness.

Flight Number Date

KA852 20 Aug 2018
KA850 20 Aug 2018
KA852 21 Aug 2018
KA770 21 Aug 2018
Surrogate Key

▪ A surrogate key is a unique, numeric value that is intentionally added to a

relation by database administrator for serving as the primary key.
▪ Surrogate key values have no meaning to users.
▪ Surrogate key are often used when there is no columns in a relation that can
naturally served as a primary key.

InstructorID InstructorName Office Phone

Number
3 Zhang, Lu RRS708 34115880
2 LIU, Yang RRS631 34112798
9 WANG, Alan AAB556 34110001

Primary Key (Surrogate key)

Surrogate Key

▪ To avoid the use of composition key

FlightID Flight Number Date

1 KA852 20 Aug 2018
2 KA850 20 Aug 2018
3 KA852 21 Aug 2018
4 KA770 21 Aug 2018

Surrogate key
Relationship Between Tables

▪ A table may be related to other tables

▪ For example
– An instructors teaches multiple courses
– A manger controls multiple projects
A Foreign Key

▪ To establish relationships, we need to implement a foreign key.

▪ A foreign key is a primary key from one table that is placed into another table.
▪ The key is called a foreign key in the table that receives the key.

COURSE_INSTRUCTOR INSTRUCTOR Primary key

InstructorID
CourseID
InstructorName
InstructorID
Office

PhoneNumber
Foreign key
 Referential Integrity 参照完整性

▪ Referential integrity states that every value of a foreign key must match a
value of an existing primary key.
Foreign key Primary key

Course ID Instructor ID
InstructorID InstructorName Office Phone Number
COMP7990 3
3 Zhang, Lu RRS709 34115880
COMP4125 3
2 LIU, Yang RRS631 34112798
COMP7820 3
COMP4115 3 9 WANG, Alan AAB556 34110001

COMP4405 2
Does InstructorID 4 exist in INSTRUCTOR table?
COMP4005 4

New record
Null Values

▪ A null value means that no data exists

– An empty cell in the table

▪ This is different from a zero, space character, empty string, or tab character

▪ The problem of null values

模棱两可的
– A null if often ambiguous. It could mean
➢ The column value is unknown
➢ The column value has not be decided yet
➢ The column value is not applicable
– Each may have entirely different implications
Functional Dependency
意味着只需要存储函数关系

▪ A relationship between attributes in which one attribute (or group of

attributes) determines the value of another attributes in the same table.

▪ Example
– The price of one cookie can determine the price of a box of 12 cookies
(cookiePrice, Qty) → BoxPrice
Candidate/Primary Keys and Functional Dependency

▪ By definition, a candidate key of a relation will functionally determine all other

non-key attributes in the row.

▪ Likewise, a primary key of a relation will functionally determine all other non-
key attributes in the row.

InstructorID → (InstructorName, Office, PhoneNumber)

Outline

▪ Introduction to Databases

▪ Relational Data Model

▪ Introduction to SQL
Structured Query Language (SQL)

▪ Structured
首字母缩略词
Query Language
– Acronym: SQL
– Pronounced “sequel” or “S-Q-L”
– Original developed by IBM as the SEQUEL language in the 1970s
– Based on relational algebra
代数
SQL

▪ SQL is not a programming language, but rather is a data sub-language

▪ SQL is comprised of
– A Data Definition Language (DDL)
➢ Used to define and manage database structures
– A Data Manipulation Language (DML)
➢ Data definition and updating
检索
➢ Data retrieval (Queries)
– A Data Control Language (DCL)
➢ For creating user accounts, managing permissions, etc
SQL for Data Definition

▪ The SQL data definition statements includes

– CREATE
➢ To create database objects
– ALTER
➢ To modify the structure and/or characteristics of existing database objects
– DROP
➢ To delete existing database objects
SQL for Data Definition: CREATE

▪ Creating database tables

– The SQL CREATE TABLE statement
name of the table

除了在命令窗口中，分号；不是必须的，空格也不是必须的
CREATE TABLE Employee ( 关键字的大小写不敏感
empId Integer NOT NULL, case insensitive
case sensitive
empName Char(25) NOT NULL
);

data type of the column

name of the column
SQL for Data Definition: CREATE with CONSTRAINT

▪ Creating database tables with PRIMARY KEY constraints

– The SQL CREATE TABLE statement
– The SQL CONSTRAINT keyword

CREATE TABLE Employee (

empId Integer NOT NULL,
empName Char(25) max25 NOT NULL,
CONSTRAINT empPk PRIMARY KEY(empId)
); name of constraint
SQL for Data Definition: CREATE with CONSTRAINT

▪ Creating database tables with composite primary keys using PRIMARY KEY
constraints
– The SQL CREATE TABLE statement
– The SQL CONSTRAINT keyword

CREATE TABLE EmplyeeSkill (

empId Integer NOT NULL,
skillId Integer NOT NULL,
skillLevel Integer NULL,
CONSTRAINT empSkillPk PRIMARY KEY(empId, skillId)
);
SQL for Data Definition: CREATE with CONSTRAINT

▪ Creating database tables using PRIMARY KEY and FOREIGN KEY

constraints
– The SQL CREATE TABLE statement
– The SQL CONSTRAINT keyword

CREATE TABLE EmplyeeSkill (

empId Integer NOT NULL,
skillId Integer NOT NULL,
skillLevel Integer NULL,
CONSTRAINT empSkillPk PRIMARY KEY(empId, skillId),
CONSTRAINT empFk FOREIGN KEY(empId) REFERENCES Employee(empId),
CONSTRAINT skillFk FOREIGN KEY(skillId) REFERENCES Skill(skillId)
);
 Employee EmployeeSkill Skill

empId empName empId skillId skillLevel skillId skillName

1 Dan 1 101 5 101 C

2 Penny 1 102 3 102 Java

2 101 4
3 Sheldon 103 Python

2 103 5

3 102 3

3 102 5 foreign key in italic

primary key in bold and
underscore
SQL for Data Definition: CREATE with CONSTRAINT

▪ Creating database tables using PRIMARY KEY and FOREIGN KEY

constraints
– The SQL CREATE TABLE statement
– The SQL CONSTRAINT keyword
如果不设置cascade，DBMS不会允许更改操作，
– ON UPDATE CASCADE and ON DELETE CASCADE 因为正在其他table中被使用

CREATE TABLE EmplyeeSkill (

empId Integer NOT NULL,
skillId Integer NOT NULL,
skillLevel Integer NULL,
CONSTRAINT empSkillPk PRIMARY KEY(empId, skillId),
CONSTRAINT empFk FOREIGN KEY(empId)
REFERENCES Employee(empId) ON DELETE CASCADE,
CONSTRAINT skillFk FOREIGN KEY(skillId)
REFERENCES Skill(skillId) ON UPDATE CASCADE
);
 Cascade Delete

Employee EmployeeSkill Skill

empId empName empId skillId SkillLevel skillId skillName

1 Dan 1 101 5
101 C
2 Penny 1 102 3
102 Java
3 Sheldon 2 101 4
103 Python
2 103 5

3 102 3
foreign key in italic
3 102 5 primary key in bold and
underscore
When we delete a record (e.g. employee 1) from Employee table, “cascade delete” will
automatically delete related records from other related tables.
Cascade Update

Employee EmployeeSkill Skill

empId empName empId skillId SkillLevel skillId skillName

1 101
105 5
1 Dan 105
101 C
1 102 3
2 Penny 102 Java
2 101
105 4
3 Sheldon 103 Python
2 103 5

3 102 3 foreign key in italic

primary key in bold and
3 102 5
underscore
When we update the skillId for ‘101’ to ‘105’ in skill table, “cascade update” will automatically update
related records in other related tables.
 Quick Question

Employee EmployeeSkill Skill

empId empName empId skillId SkillLevel skillId skillName

1 Dan 1 101 5
101 C
2 Penny 1 102 3
102 Java
3 Sheldon 2 101 4
103 Python
2 103 5

3 102 3
foreign key in italic
3 102 5 primary key in bold and
underscore
What happen if we delete that data? Any cascade update/delete?
No
Primary Key Constraint: ALTER

▪ Adding primary key constraints to an existing table

– The SQL ALTER statement

ALTER TABLE Emplyee

ADD CONSTRAINT empPk PRIMARY KEY(empId);
Employee

PRIMARY KEY empId empName

1 Dan

2 Penny

3 Sheldon
Composite Primary Key Constraints: ALTER

▪ Adding a composite primary key constraint to an existing table

– The SQL ALTER statement
ALTER TABLE EmplyeeSkill EmployeeSkill
ADD CONSTRAINT empSkillPk
PRIMARY KEY(empId, skillId);
empId skillId SkillLevel
PRIMARY KEY 1 101 5

1 102 3

2 101 4

2 103 5

3 102 3

3 102 5
Foreign Key Constraint: ALTER

▪ Adding foreign key constraints to an exisiting table

– The SQL ALTER statement

ALTER TABLE EmplyeeSkill

ADD CONSTRAINT empFk FOREIGN KEY(empId)
REFERENCES Employee(empId)
Modifying Data using SQL

▪ INSERT INTO
– Add a new row to a table
▪ UPDATE
– Update the rows in a table which match the specified criteria
▪ DELETE FROM
– Delete the rows in a table which match the specified criteria
Adding Data: INSERT INTO

▪ To add a new row to an existing table

▪ Non-numeric data must be enclosed in single quotes (‘’)
INSERT INTO Employee (empId, salaryCode, lastName)
VALUES (62, 11, 'Halpert');

Employee Employee

empId salaryCode lastName empId salaryCode lastName

1 2 Brown 1 2 Brown

22 7 Smith 22 7 Smith

62 11 Halpert
Quick Question

▪ How to insert the record for "Kevin" with salaryCode 3, his empId is 10
▪ Which of the following records can be insert successfully?
• "Brown", salaryCode 2, salary 23000, empId 3
• "Steve", salaryCode 5, salary 30000, empId 4
• "Brady", salaryCode 7, empId 22
Salary Employee

empId salaryCode salary empId salaryCode lastName

2 23000 1 2 Brown

3 26000 22 7 Smith

7 35000 62 11 Halpert
Changing Data Values: UPDATE
▪ To change the data values in an existing row (or a set of rows) use the
UPDATE statement 匹配或过滤标准
▪ The WHERE clause specifies the matching or filtering criteria for the records
(rows) that are to be displayed
UPDATE Employee Update the phone number to
SET phone = '34111001' ‘34111001’ for the employee
WHERE empId = 29; whose empID is 29

Employee Employee

empId phone empId phone

18 69981245 18 69981245

29 12345678 29 34111001
Changing Data Values: UPDATE
▪ To change the data values in an existing row (or a set of rows) use the
UPDATE statement
▪ The WHERE clause specifies the matching or filtering criteria for the records
(rows) that are to be displayed
UPDATE Employee Update the deptID to
SET deptId = 4 4 for the employee
WHERE empName LIKE 'Da%'; names starts with "Da"
Employee Employee

deptId empName deptId empName

WHERE, LIKE are 18 Dave 4 Dave

some keywords that
will be covered more 29 John 29 John
next week!

29 Dan 4 Dan
Quick Question
▪ What if we say
UPDATE Employee
SET deptId = 5
WHERE empName Like '%k'
▪ And what if we say
UPDATE Employee
SET deptId = 5

Employee

deptId empName

WHERE, LIKE are 18 Dave

some keywords that
will be covered more 29 John
next week!

29 Dan
Deleting Data: DELETE

▪ To delete a row (or a set of rows) from a table use the DELETE FROM
statement

DELETE FROM Employee Employee Employee

WHERE empId = 29;
empId phone empId phone
DELETE FROM Employee
WHERE empName LIKE 'Da%'; 18 69981245 18 69981245

29 34110000
DELETE FROM Employee;
Next Week:

▪ SQL Query
• SELECT
• WHERE
• GROUP BY
• HAVING
• join tables