G4G0-1/Semester 1/Database Systems/Week 10/Week 10 Database Systems.md

Lecture 1 (13:00) - Relational Algebra

Data Manipulation Language

• DML is a language providing a set of operations on the data of the database
• Part of DML that involves data retrieval is called query language. One example is SQL

#### Relational Calculus

• Non-procedural language - declarative.
• Used to formulate definition of relation in terms of one or more relations.
• Specifies what is to be retrieved, rather than how.

Relational Algebra

• High level procedural language
• Used to tell DBMS how to build new relation from one or more relations in database

Fundamental Operations

SQL Overview

Selection

U+03C3 <selection condition> (<relation name>)

• Selects a subset of the tuples in a relation that satisfy a selection condition.
• sigma (lc) is used to specify a selection.
• Selection condition is a bool expression, that includes AND, NOT, OR.
• Result is a relation that has the same attributes as the relation that with the name specified.
• ex.

Projection

• Selects a subset of the attributes of a relation.
• Pi (uc) is used to define a projection
• Order of attributes in the result is the same as that in the attribute list
• Duplicate tuples are removed.
• ex.

Closure

• Operations work on one or more relations to define another relation, without changing the original.
• Both operands and results are relations.
• Output from one operation can become the input from another.
• Parallels algebraic operations in arithmetic, taking one or more numbers as operands, returning a number as output.
• ex.
  • Select all staff with salary greater than 20000, and exclusively output the name.
  • Both input and output are relations.

Unary and Binary Operations

Unary Operations

• Operate on one relation
• Selection
• Projection

Binary Operations

• Operate on two relations
• Cartesian products
• Union
• Set difference

Union

• Union of 2 relations is a relation that contains all tuples of both.
• Duplicates removed.
• Both relations must be union compatible.
• Resulting relation might have the same attributes as 1st or 2nd relation.
• ex.

Union Compatibility

• Two relations R1(A1, A2,…. An) and R2(B1, B2,…. Bm) are union compatible if, and only if n = m and domain(Ai) = domain(Bi) for 1 =< i =< n.
• It is rare that two relations are union compatible
• Projection may be used to make two relations union compatible.

Union with Projection

• This is also an example of closure, as the output for the projection is used as the input for the union.

Set Difference

• In two relations, the set difference includes all tuples that are in one, but not the other.
• Resulting relation may have the same attribute names as the first or second relation.
• Must be union compatible.
• ex.

Cartesian Product

• Relation which has concatenation of every tuple of first relation with every relation of the second.
• If the 2 relations have attributes with the same name, attribute names are prefixed with the relations name.
• ex.

Tutorial 2 (15:00) - Relational Algebra

1. List the names and types of all boats

• Relational Algebra Pi boat-name, boat-type (boats)

SELECT DISTINCT boat-name, boat-type
FROM boats;

1. List the names of all pirates Pi name (Pirates)

SELECT DISTINCT p-name
FROM pirates

1. List the names of all pirates who have less than two eyes Pi p-name ( Sigma eyes < 2 (pirates) )

SELECT DISTINCT names
FROM pirates
WHERE eyes < 2;

1. List the names of all pirates who have less than two eyes or less than two legs or both. Pi p-name ( Sigma eyes < 2 OR legs < 2 (pirates) )

SELECT DISTINCT names
FROM pirates
WHERE eyes < 2 OR legs < 2;

1. List all the places that either pirates or boats or both originate from Pi origin (pirates) U Pi origin (boats)

SELECT origin FROM pirates
UNION
SELECT origin FROM boats;

• NOTE: order matters, but only to the order in which they would appear in the list.

1. List all the places where boats but not pirates originate from Pi origin (boats) - Pi origin (pirates)

SELECT origin FROM boats
EXCEPT
SELECT origin FROM pirates;

• NOTE: This specifically removes all places pirates originate from from the selection.

Workshop 1 (10:00)

• Implement the two tables using the SQL CREATE TABLE command.

CREATE TABLE employee(
	empNo INTEGER PRIMARY KEY
	eName VARCHAR(20)
	sal INTEGER
);

CREATE TABLE manager(
	empNo INTEGER PRIMARY KEY
	eName VARCHAR(20)
	sal INTEGER
);

• Write SQL statements that populate the two tables with the data shown.

INSERT INTO employee
VALUES (7782, "clark", 2450), (7839, "king", 5000);

INSERT INTO manager
VALUES (8034, "smith", 7000),(8044, "yao", 6000);

• Write a single SQL query which returns all the staff from both of the tables.

SELECT empNo, eName
FROM employee
INNER JOIN manager
ON employee.empNo = manager.empNo;