comp 430 intro. to database systemsphys 101 1 baker phys 102 2 baker response counter. what is key?...

COMP 430Intro. to Database Systems

Single-table SQL

Slides use ideas from Chris Ré and Chris Jermaine.

Get clickers today!

SELECT name FROM sqlite_master WHERE type='table'

Clicker test – Have you used clickers before?

A. True

B. False

True

False

0%0%

What is Structured Query Language (SQL)?

“Standard” language for databases• Many related standards: ANSI SQL, SQL92 (= SQL2), SQL99 (= SQL3)• Vendors support various subsets and extensions

Very high-level programming language – Highly optimized, parallelized

Huge language – We’ll cover only a core subset.

Typically called from main application language (Python, Java, …)• SQLexecute(“…SQL CODE…”)

Two languages in one

• Data Definition Language (DDL)• Define relational schemas declarative

• Create/modify/delete tables and their attributes imperative

• Data Manipulation Language (DML)• Query DB for desired data declarative

• Add/modify/delete data in tables imperative

• Define/use functions, procedures, triggers imperative

Basic concepts

Table=Relation

p_name price manufacturer

Gizmo $19.99 GizmoWorks

Powergizmo $39.99 GizmoWorks

Widget $19.99 WidgetsRUs

HyperWidget $203.99 Hyper

Product

Column=attribute=field:A typed data entry in each row.

#columns = arity=degree

Row=tuple=record:A single entry in the table.#rows = cardinality

Heading:Names of each column.

Table=Relation

Product

Table is a multiset of rows. Not ordered. Duplicate rows allowed.






Schema defines table attributes

Product

Product (p_name: string, price: float, manufacturer: string)






Creating a table


CREATE TABLE Product (p_name VARCHAR(50),price NUMERIC(6,2),manufacturer VARCHAR(50)

);

Some SQL variants have a CURRENCY or

MONEY type.

SQL attribute types

All types are atomic!

Examples:• CHAR(n), VARCHAR(n), NCHAR(n), NVARCHAR(n)

• INT, BIGINT, SMALLINT, FLOAT, DECIMAL(m,n)

• BOOLEAN, MONEY, DATETIME

Traditionally. However, sets and arrays are allowed

in some SQL versions.

NULL data

Any value of any type can be NULL.• Signals missing value.

• Value doesn’t exist.

• Value exists but unknown.

• Value not applicable to this record.

Keys

Key = minimal set of attributes acting as a unique tuple identifier.• Consider the universe of all potential relation data, not just what the table

currently contains.

Product


Gizmo 19.99 GizmoWorks

Powergizmo 39.99 GizmoWorks

Widget 19.99 WidgetsRUs

HyperWidget 203.99 Hyper

What is key?

A. dept

B. dept, number

C. dept, number, section

D. dept, number, section, instructor

dept

dept, num

ber

dept, num

ber, s

ectio

n

dept, num

ber, s

ectio

n, inst

ruct

or

25% 25%25%25%

Course

dept number section instructor

MATH 101 1 Jones

MATH 101 2 Smith

MATH 102 1 Williams

PHYS 101 1 Baker

PHYS 102 2 Baker

Response Counter

What is key?

A. s_name

B. abbrev

C. The pair s_name, abbrev

D. Each of s_name and abbrev

s_nam

e

abbre

v

The p

air s

_name, a

bbrev

Each

of s

_nam

e an

d abbre

v

25% 25%25%25%

s_name abbrev year_admitted

Alabama AL 1819

Alaska AK 1959

Arizona AZ 1912

Arkansas AR 1836

California CA 1850

State

Response Counter

Primary keys

Every table should have one primary key.• Each tuple must have distinct non-NULL values of primary key attributes.

• Guarantees table is a mathematical relation.

Product







Primary keys are a DB-checked constraint

Adding a new record with duplicate or NULL primary key will fail.• Failure reported as exception or erroneous return value.

Product






Widget 15.99 GizmoWorks

We will see other kinds of DB-checked constraints.


Creating a table with primary key


CREATE TABLE Product (p_name VARCHAR(50),price DECIMAL(6,2),manufacturer VARCHAR(50),PRIMARY KEY (p_name)

);

NOT NULL constraint

Course (c_id: integer, c_name: string, instructor: string)

CREATE TABLE Course (c_id INTEGER,c_name VARCHAR(50) NOT NULL,instructor VARCHAR(50),PRIMARY KEY (c_id)

);

Adding a single record to a table

Product






INSERT INTO Product VALUES (‘MiniWidget’, 21.99, ‘WidgetsRUs’);INSERT INTO Product (p_name, manufacturer)VALUES (‘NanoWidget’, ‘WidgetsRUs’);

Product






MiniWidget 21.99 WidgetsRUs

NanoWidget NULL WidgetsRUs

Activity: Creating a table

02a-table.ipynb

Course

dept number section instructor

MATH 101 1 Jones

MATH 101 2 Smith

MATH 102 1 Williams

PHYS 101 1 Baker

PHYS 102 2 Baker

Course (dept, number, section, instructor)

CREATE TABLE Course (…);INSERT INTO Course …;

Activity partial solution

Course (dept, number, section, instructor)

CREATE TABLE Course (dept CHAR(4),number CHAR(3),section INT DEFAULT 1,instructor VARCHAR(50),PRIMARY KEY (dept, number, section)

);

Convenient for this example.

Simple queries

SELECT-FROM-WHERE

SELECT attributesFROM tablesWHERE conditions

Results in a new table, which can be returned or stored.

Selecting some records

Product






SELECT *FROM ProductWHERE Price > 20;




More selection examples

SELECT *FROM ProductWHERE price IS NOT NULL;

SELECT *FROM ProductWHERE price > 20 AND manufacturer = ‘GizmoWorks’;

SELECT *FROM ProductWHERE manufacturer IN (‘GizmoWorks’, ‘WidgetsRUs’);

SELECT *FROM ProductWHERE price BETWEEN 20 AND 40;

Selection with pattern matching

SELECT *FROM ProductWHERE p_name LIKE ‘%Gizmo%’;

% = Match any sequence of 0-or-more characters_ = Match any single character[abc] = Match any one character listed[a-c] = Match any one character in range

Projecting some fields

Product






SELECT p_name, manufacturerFROM Product;

p_name manufacturer

Gizmo GizmoWorks

Powergizmo GizmoWorks

Widget WidgetsRUs

HyperWidget Hyper

Answer (p_name, manufacturer)

Combining selection & projection

Product






SELECT p_name, manufacturerFROM ProductWHERE price > 20;

p_name manufacturer


HyperWidget Hyper

Results not necessarily distinct

Product






SELECT manufacturerFROM Product;

manufacturer

GizmoWorks

GizmoWorks

WidgetsRUs

Hyper

Tables are multisets!

Answer (manufacturer)

Making results distinct

Product






SELECT DISTINCT manufacturerFROM Product;

manufacturer

GizmoWorks

WidgetsRUs

Hyper

Ensures results are a set.

Query semantics – set notation

SELECT [DISTINCT] a1, a2, …, am

FROM TWHERE Conditions(a’1, a’2, …, a’p);

{(a1, a2, …, am) | Conditions(a’1, a’2, …, a’p)}Multisets by default.Sets with DISTINCT.

Query semantics – sequence of ops

SELECT [DISTINCT] a1, a2, …, am

FROM TWHERE Conditions(a’1, a’2, …, a’p);

Answer = {}for row in T do

if Conditions(row.a’1, row.a’2, …, row.a’p)then Answer = Answer {(row.a1, row.a2, …, row.am)}

return Answer

Multiset union by default.Set union with DISTINCT.

Subset of results

Product






SELECT p_name, manufacturerFROM ProductLIMIT 2;

p_name manufacturer

Gizmo GizmoWorks


Which 2 is implementation-

dependent.

Computation in SELECT clause

Product






MiniWidget 21.99 WidgetsRUs

NanoWidget NULL WidgetsRUs

SELECT p_name, IsNull(price, 0) AS priceFROM Product;

p_name price

Gizmo 19.99

Powergizmo 39.99

Widget 19.99

HyperWidget 203.99

MiniWidget 21.99

NanoWidget 0

More computation in SELECT clauses

SELECT location, time, celsius * 1.8 + 32 AS fahrenheitFROM SensorReading;

SELECT player_id, Floor(height) AS feet, (height – Floor(height)) * 12 AS inchesFROM Player;

Use AS. Without it, SQL will create a default field name.

Sorting

Product






SELECT p_name, manufacturerFROM ProductORDER BY price DESC, manufacturer;

p_name manufacturer

HyperWidget Hyper


Gizmo GizmoWorks

Widget WidgetsRUs

Sorting vs. Multiset semantics

Table rows are unordered, except when they’re ordered.

More accurately, unless you use ORDER BY:• Can’t assume anything about ordering.

• Ordering depends on implementation, which can vary.

• Queries don’t necessarily maintain order of original table.

Activity: Writing queries

02b-queries.ipynb

Some details

NULL semantics

NULL is not a value. It is the lack of a value.

In numeric operations:• f(NULL) NULL

In Boolean operations, we use 3-value logic (FALSE, UNKNOWN, TRUE):• NULL = ‘Houston’ UNKNOWN

• NULL = NULL UNKNOWNA NOT A

F T

U U

T F

A AND BB

F U T

A

F F F F

U F U U

T F U T

A OR BB

F U T

A

F F U T

U U U T

T T T T

Syntax details

• Strings use single quotes, not double ‘Houston’

• Equality test uses single =, not double x = 5

• Amount of whitespace doesn’t matter

Syntax details – case sensitivity

• Case insensitive• Keywords: SELECT, NULL• Table names: Product• Function names: IsNull()

• Depends on SQL version• Attribute names: product_id

• Case sensitive• String literals: ‘HOUSTON’, ‘Houston’, ‘houston’

comp 430 intro. to database systemsphys 101 1 baker phys 102 2 baker response counter. what is key?...

Documents