Phone books: a database?

• It contains formated information
• Information is indexed
• It's possible to find a piece of information (phone number, address) from a name

Phone book: a bad database!

• Time consuming: it contains a large number of entries
• Indexed by last/first names
• Once it's printed, there is no more updates

Hierarchical database system

Network database system

Relational model

Entity

Something of interest to the database user community. Examples include customers, parts, geographic locations, etc.

Column

An individual piece of data stored in a table.

Row

A set of columns that together completely describe an entity or some action on an entity. Also called a record or a tuple.

Table

A set of rows, held either in memory (nonpersistent) or on permanent storage (persistent).

Result set

Another name for a nonpersistent table, generally the result of an SQL query.

Primary key

One or more columns that can be used as a unique identifier for each row in a table.

Foreign key

One or more columns that can be used together to identify a single row in another table.

SQL statements

• Schema statement to define the data structures stored in the database
• Data statement to manipulate the data
• Transaction statement to begin, end, and roll back transactions

A progamming language

• Nonprocedural
• Dependant to other environments

A server

From your laptop!

Or in the cloud

Cloud9

Connect

Create a workspace

Presentation of C9

Setup PostgreSQL

Copy this code in the CLI

$ sudo service postgresql start
$ sudo sudo -u postgres psql

First interaction

Try it!

-- Get the current date
SELECT CURRENT_DATE;
/**
 *     date    
 * ------------
 *  2016-02-16
 * (1 row)
 */

How to work with C9

1. Create a new folder sql in your File exporer
2. Inside this folder create a first script 01-awesome-name.sql
3. Edit this file in the IDE
4. SELECT CURRENT_DATE;
5. Save: cmd + s or ctrl + s or File > Save
6. Run the script from the CLI, in psql
7. \i sql/01-awesome-name.sql

How to work: result

Help

Open PSQL

$ sudo sudo -u postgres psql

Run a script

postgres=# \i PATH/to-script.sql

Close the client PSQL

postgres=# \q

One rule

All the values of a column have to be of the same type

Types, types, types

• boolean
• character varying [ (n) ]
• text
• date
• integer
• double precision
• ...

PostgreSQL data types

I want to store some information about the students of GA

I create my first table Student

How do I describe a student?

How do I describe a student?

• First name
• Last name
• Email
• Age
• ...

I create my first table Student

Now we can insert some rows

Let's integrate the cities

Let's integrate the classes

How to link Students and Cities?

A Student has only one City

SQL Constraint

Pattern A

Add a Foreign Key in the Student table referring to the City table

SQL Constraint

Pattern A

1. Create Primary keys in all the tables
2. Create a Foreign key in the Student table

SQL Constraint

Pattern A

SQL Constraint

Pattern A

How to link Students and Classes?

A Student has one or many Classes

SQL Constraint

Pattern B

Create a pivot table with 2 Foreign Keys referring to Student table and to Class table

The Foreign key couple is in fact the Primary key of this pivot table

SQL Constraint

Pattern A

Query sample

-- Sample query
select o.order_id "Order ID", count(distinct i.id) "Quantity" from orders o inner join items i on o.order_id = i.order_id where o.delivered_at is not null and not i.is_sample group by o.order_id order by 1 desc limit 100;

Please, indent

-- Indented sample query
SELECT
    o.order_id           AS "Order ID",
    COUNT(DISTINCT i.id) AS "Quantity"
FROM
    orders o
INNER JOIN
    items i
        ON o.order_id = i.order_id
WHERE
    o.delivered_at IS NOT NULL AND NOT i.is_sample
GROUP BY
    o.order_id
ORDER BY o.created_at DESC
LIMIT 100;

Not only for you, but also for everyone else!

Query 1

Create a table

-- Create table query
DROP TABLE IF EXISTS student;
CREATE TABLE student (
    id SERIAL PRIMARY KEY,
    first_name VARCHAR(255),
    last_name VARCHAR(255),
    email VARCHAR(255),
    age INTEGER
);

Documentation

Query 2

Insert some columns

-- Insert rows
INSERT INTO student (
    first_name,
    last_name,
    email,
    age
) VALUES (
    'Leo',
    'Di Caprio',
    'leo@gmail.com',
    40
), (
    'Angelina',
    'Joli',
    'angi@gmail.com',
    42
);

Documentation

Query 3

Retrieve your data

-- Select rows
SELECT
    id,
    first_name,
    last_name
FROM
    student
LIMIT 10;

Documentation

Your turn!

Create the tables City and Class and insert some data inside

Query 4

See your table

\d+ city

Query 5

Update a table

-- Update a table
ALTER TABLE student
    ADD city_id INTEGER REFERENCES city(id);

Documentation

Query 6

Update rows

-- Update rows
UPDATE
    student -- Table name
SET
    city_id = 1 -- Field to update
WHERE
    id = 1 -- Row to update
;

Documentation

Your turn!

Create the pivot table student_class

Query 7

Retrieve you data (bis)

-- Select some data
SELECT
    s.first_name AS "Student first name",
    s.last_name AS "Student last name",
    c.name AS "City"
FROM
    student s
    INNER JOIN city c ON c.id = s.city_id
WHERE
    c.name = 'NYC'
ORDER BY
    s.first_name DESC
LIMIT 10;

Documentation

Set up the dataset

# Leave PSQL if opened with \q
$ wget https://raw.githubusercontent.com/aeud/sql-pres/master/movielens.sql # Download the dataset
$ sudo sudo -u postgres psql < movielens.sql # Add the dataset to your database
$ sudo sudo -u postgres psql # Restart PSQL client

You just created 6 tables

• occupations
• users
• ratings
• movies
• genres
• genres_movies

Relations

A User has one Occupation

A Rating has one User and one Movie

A Movie has many Genres (via genre_movie)

Fields

-- PostgreSQL
\d+ occupations
\d+ users
\d+ ratings
\d+ movies
\d+ genres
\d+ genres_movies

Sample

-- Indented sample query
SELECT
    o.order_id           AS "Order ID",
    COUNT(DISTINCT i.id) AS "Quantity"
FROM
    orders o
INNER JOIN
    items i
        ON o.order_id = i.order_id
WHERE
    o.delivered_at IS NOT NULL AND NOT i.is_sample
GROUP BY
    o.order_id
ORDER BY o.created_at DESC
LIMIT 100;

Query clauses

SELECT

-- Apply these queries
SELECT * FROM users LIMIT 5;
SELECT id, age, gender FROM users LIMIT 5;
SELECT id AS "User ID", age AS "Age", gender AS "Gender" FROM users LIMIT 5;

FROM

-- Apply these queries
SELECT * FROM users LIMIT 5;
SELECT * FROM movies LIMIT 5;

WHERE

-- Apply these queries
SELECT * FROM users WHERE id = 1 LIMIT 5;
SELECT * FROM users WHERE gender = 'M' LIMIT 5;

GROUP BY

Use case

How to determine how many females / males are recorded in our dataset?

Aggregate functions

• SUM(field)
• COUNT(field)
• COUNT(DISTINCT field)

Documentation

GROUP BY

-- Apply this queries
SELECT
    u.gender AS "Gender", -- Dimension(s)
    count(u.*) AS "Count" -- Aggregate function
FROM
    users AS u
GROUP BY
    gender -- Dimension(s)
;

ORDER BY

-- Apply these queries
SELECT * FROM users ORDER BY age ASC LIMIT 5;
SELECT * FROM users ORDER BY age DESC LIMIT 5;
SELECT * FROM users ORDER BY gender ASC LIMIT 5;

LIMIT

-- Apply these queries
SELECT * FROM users ORDER BY age ASC LIMIT 5;
SELECT * FROM users ORDER BY age ASC LIMIT 10;
SELECT * FROM users ORDER BY age ASC LIMIT 100; -- use \q to escape

Exercise

How can I get the distribution of the movies, by release year?

Conditions

The WHERE clause is made of a serie of expressions, linked by logical operators

Logical operators

• AND
• OR
• NOT

Rules

• TRUE AND FALSE = FALSE
• TRUE OR FALSE = TRUE
• TRUE AND (FALSE OR TRUE) = TRUE
• TRUE AND NOT(FALSE OR TRUE) = FALSE

Expressions

• Number
• Column in a table
• Built-in functions (Documentation)
• Comparaison operators: =, !=, <, >, <>, LIKE, IN
• Arithmetic operators, such as +, −, *, /

SQL

-- Conditions
SELECT
    id,
    age,
    gender
FROM
    users u
WHERE
    u.gender = 'M' AND u.age < 35
ORDER BY
    id desc
LIMIT 5;

Your turn

With the documentation, write a query to fetch the movies released in 1996 with a title starting with the letter D or T

Your turn (Bonus)

Get the distribution of movies released in 1996, by title first letter

Get the distribution of movies with a title starting with P, by year of release

How to get how many time a movie has been rated? And by which user?

Join

-- Join, use aliases!!
SELECT
    m.title, -- 1
    count(r.id) reviews, -- 2
    count(distinct r.user_id) unique_reviews -- 3
FROM
    movies m
    INNER JOIN ratings r ON r.movie_id = m.id
ORDER BY
    2 desc
LIMIT 5;

Join

-- Join, use aliases!!
SELECT
    m.title, -- 1
    count(r.id) reviews, -- 2
    count(distinct r.user_id) unique_reviews -- 3
FROM
    movies m
    INNER JOIN ratings r ON r.movie_id = m.id
GROUP BY
    1
ORDER BY
    2 desc
LIMIT 5;

Your turn

Get the top 10 movies, by average rate

Your turn (Bonus)

Get some statistics by year, about the number of rates, the average rates, etc.. Be creative

Try to find an answer to this question: Does users give rates to movies only if they liked it? only if they hated it? only if they liked or hated it? (tip: analyse the distribution of rates)

Different ways to join