Page History

Introduction

What is

SQL (Structured Query Language) is the standard language for managing data in relational database managements systems (DBMS).  It can be used to Create, Read, Update, and Delete data within the databases.

What you'll learn:

You will learn the basic syntax and practice writing SQL queries.

Prerequisites:

Basic understanding of relational or tabular data.

Tutorial:

a Database

A database is a structured collection of data. It is designed to efficiently store, retrieve, and manage information.

The most widely used type of databases are relational databases (PostgresSQL, MySQL, SQL Server, Oracle, etc.).  In a relational database, data is stored in tables (rows and columns) and the tables can be have relationships with other tables.  For example, a Patient Demographic table will be related to a Patient Medication table through the Patient ID.

About SQL

 'SQL' (Structured Query Language) is the language used to manipulate data is relational databases.  SQL can be used to Create, Read, Update, and Delete data within relational databases.

Databases that use SQL

• SQL Server
• Oracle
• Postgres SQL
• MySQL
• DB2
• ... and more

Other types of databases besides relational databases:

• NoSQL
  • document database
  • graph database
  • key-value
  • time series

However a lot of NoSQL databases have incorporated SQL-like syntax so that users can query their databases.

GUI Tool for SQL

• DBeaver (Paid and Free editions) https://dbeaver.io/download/

Tutorial Prerequisites

Basic understanding of tabular data (like Excel).

Project Setup

We will use the following site to test SQL commands: https://sqliteonline.com/

Connect to the 'PostgreSQL' database.  Note that this site has a 15 minute idle limit.

Commands to insert data

These commands are also explained in later sections of the tutorial.  For teaching purposes, they have been replicated for project set up so that users can learn how to query data first (i.e. the most likely step that users are performing).

-- Create the following two

Step 0 - Project Setup

We will use the following site to test SQL commands: https://sqliteonline.com/

Connect to the 'PostgreSQL' database.  Note that this site has a 15 minute idle limit.

Step 1 - Create Tables for Tutorial

Person Table

Country Table

The two red columns (person.country_id and country.id) is meant to indicate a foreign-key relationship.  For example, it can be read as: "Sally lives in country USA".

-- Create the following two tables for this demo:
CREATE TABLE personcountry ( 
	id INTEGER PRIMARY KEY, person
	country_name VARCHAR(100), income NUMERIC(7,2)  
);

CREATE TABLE countryperson ( 
	id INTEGER PRIMARY KEY, country
	person_name VARCHAR(100), 
	income NUMERIC(7,2), 
	country_id integer,
	CONSTRAINT country_id_fk
		FOREIGN KEY(country_id) 
		REFERENCES );

For more information on creating tables: https://www.postgresql.org/docs/9.2/sql-createtable.html

Step 2 - Insert Data into the Tables

--country(id)
);

-- Insert data into the 'personcountry' table:
INSERT INTO personcountry (id, personcountry_name, income) VALUES (1, 'SallyCanada'), 60000);
INSERT (2, 'USA'), (3, 'Mexico'), (4, 'Sweden');

-- Insert data into the 'person' table:
INSERT INTO person (id, person_name, income, country_id) VALUES (21, 'BobSally', 60000, 700001);
INSERT INTO person (id, person_name, income, country_id) VALUES (32, 'LucyBob', 70000, 800001);

Insert data into the 'country' table:

INSERT INTO countryperson (id, person_name, income, country_nameid) VALUES (13, 'CanadaLucy'), (280000, 'USA'),2);
INSERT INTO person (3, 'Mexico');

Step 3 - Update

Update the person table to have a column for the person's country:

ALTER TABLE person ADD column country_id integer;

Update the data:

UPDATE person
SET country_id = 2
WHERE person_name = 'Lucy';

UPDATE person
SET country_id = 1
WHERE person_name in ('Sally', 'Bob');

id, person_name, income, country_id) VALUES (4, 'Bill', 75000, NULL);

Result - Person Table

Result - Country Table

The two red columns (person.country_id and country.id) is meant to indicate a foreign-key relationship.  For example, it can be read as: "Sally lives in country USA".  There are different benefits for having the information in separate tables but the main reason is to reduce redundant data.

Basic Data Querying

Select Statement

-- select all columns from person table

select 
	* 
from 
	person;

-- we can also specify the columns we want to query by replacing the asterisk with the column name

select 
	id, 
	person_name, 
	income 
from 
	person;

'Order by' clause

-- if we want to return the data and have it ordered according to a specific column(s), we can use the order by clause

select * from person
order by income asc;

select * from person
order by income desc;

'Where' Clause

-- filter data from person table by name
select * from person where person_name = 'Lucy';

-- sometimes we don't know the exact spelling or word.  In that case we can use a wildcard
select * from person where person_name like '%Lu%';

'Distinct' Clause

-- get the unique countries that a person lives in

select distinct country_id from person;

Join Types

Join statements are a way to combine data from multiple tables.  In a database, you'll often find that data is spread out across multiple tables ranging from dozens to even thousands of tables.  There are multiple reasons database developers create multiple tables but the two main reasons are: make some process more efficient (inserting, querying or updating data) or it's just logically appropriate to store data in a separate table.  For example we have information about a person in the Person table and information about a country in the Country table. 

But what if we wanted to know the average income of people in Canada?  For this, we may need to join tables.

There are 4 types of joins that are available in SQL.  The following is a good diagram that explains the results from each join type:  The diagram and more information can be found here: https://www.w3schools.com/sql/sql_join.asp

Image Added

Inner Join

select 
	* 
from 
	person 
inner join country 
	on person.country_id = country.id;

Return records that are matching in both person and country table.

Left Join

select 
	* 
from 
	person 
left join 
	country 
	on person.country_id = country.id;

Returns all records from the person table and only the matching records from the country table.

Right Join

--now do a right join
select * 
from 
	person 
right join 
	country 
	on person.country_id = country.id;

Returns only matching records from the person table and all records from the country table.

Full Outer Join

select * 
from 
	person 
full outer join 
	country 
	on person.country_id = country.id;

Returns all records from person and country table.

Group By and Aggregation

Select with Group By clause

The 'GROUP BY' clause in SQL is used when you want to group rows that have the same values in certain columns and perform some kind of aggregate function (avg, max, min, count) on each group. This is particularly useful when analyzing and summarizing data.  Common usage scenarios include:

• calculating summaries
• data categorization
• reporting and visualization

For example, here we are asking the database to return the count of people per country.

select 
	country_name, 
	count(*) 
from 
	person
inner join 
	country 
	on person.country_id = country.id
group by 
	country_name;

Having Clause

Using the 'having' clause, we can take the result of the group by statement and filter it to only return countries where the number of people is greater than 1.

select 
	country_name, 
	count(*) 
from 
	person
inner join 
	country 
		on person.country_id = country.id
group by 
	country_name
having 
	count(*) > 1;

Average

Return the average income per person per country.

select 
	country_name, 
	avg(income) 
from 
	person
inner join 
	country 
		on person.country_id = country.id
group by 
	country_name;

80000.000000000000

65000.000000000000

Maximum Value

select 
	country_name, 
	'max' as agg_type, 
	max(income) 
from 
	person
inner join 
	country 
	on person.country_id = country.id
group by 
	country_name;

80000.00

70000.00

Minimum Value

select 
	country_name, 
	'min' as agg_type, 
	min(income) 
from 
	person
inner join 
	country 
		on person.country_id = country.id
group by 
	country_name;

80000.00

60000.00

Avg, Max, and Min Values:

select 
	country_name, 
	min(income), 
	max(income), 
	avg(income) 
from 
	person
inner join 
	country 
		on person.country_id = country.id
group by 
	country_name;

Union Clause

Instead of giving the queries one at a time to get the avg, max, and min, we can use a union clause.  Union clause is used to combine the results of multiple queries.  The rules for using a union clause are as follows (see link):

• Each SELECT statement within UNION must have the same number of columns
• The columns must also have similar data types
• The columns in each SELECT statement must also be in the same order

select country_name, 'avg' as agg_type, avg(income) from person
inner join country on person.country_id = country.id
group by country_name

UNION

select country_name, 'max' as agg_type, max(income) from person
inner join country on person.country_id = country.id
group by country_name

UNION
select country_name, 'min' as agg_type, min(income) from person
inner join country on person.country_id = country.id
group by country_name;

70000.00

Sub-queries

Sub-queries can be used to create intermediary tables that are then joined within the larger dataset.  It can help with applying filter logic and performance tuning by filtering the amount of data being processed in a join condition.

select * from person
full outer join (
	select * from country
	where id = 2) country_table
on person.country_id = country_table.id;

Compare the sub-query to filtering after joining:

select * from person
full outer join country
on person.country_id = country.id
where country.id = 2;

When you filter after, it takes the results of the join condition and then applies the filter logic.  The sub-query will calculate the sub-query first and then apply the join logic.

Case Statement

SELECT person_name, CASE 
WHEN income BETWEEN 50000 and 61000 THEN 1 
WHEN income BETWEEN 62000 and 71000 THEN 2 
WHEN income BETWEEN 72000 and 81000 THEN 3 END as income_category
FROM person;

Data type conversion

The following link explains how to perform conversion of data from one format into another:  https://www.postgresql.org/docs/current/functions-formatting.html

Often times when working in a big data environment, data is dumped into tables and all the columns are treated as strings.  Therefore, if you need to perform math operations or other comparisons, you may not get the correct results.

The most common conversions are string to number or string to date/timestamp.  Examples of each are below.

Treat string as a number

If you try to add two strings, it will result in an error:

select ('100') + ('100'); -- query fails

SELECT ('100'::INTEGER) + ('200'::Integer); -- result is successfully returned as 300
SELECT ('10.1'::DECIMAL) + ('12.2':: DECIMAL)' -- result is successfully returned as 22.3
SELECT ('-10.1'::DECIMAL) + ('12.2':: DECIMAL)' -- result is successfully returned as 2.1

Treat string as a date

select ('jan 01 2021') > ('feb 01 2021'); -- incorrectly results in true because 'j' is larger than 'f' alphabetically.

select ('jan 01 2021'::DATE) > ('feb 01 2021'::DATE); -- correctly returns false
select ('jan 01 2021 08:01:45'::TIMESTAMP) > ('jan 01 2021 08:01:46'::TIMESTAMP); -- correctly returns false
select to_date('05$DEC$2000', 'DD$MON$YYYY') -- returns 2000-12-05T00:00:00.000Z

Run query against table

create table test_data_conv (income_1 varchar(100), income_2 varchar(100), date_1 varchar(100), date_2 varchar(100));

INSERT INTO test_data_conv 
(income_1, income_2, date_1, date_2) 
values 
('10.1', '12.2', 'jan 01 2021', 'feb 01 2021'),
('13.1', '14.2', 'mar 01 2021', 'aug 01 2021');

SELECT sum(income_1::DECIMAL) from test_data_conv; --23.2

select * from test_data_conv where (date_1::DATE) < (date_2::DATE); -- correctly returns both rows.

ALTER TABLE person ADD column bday text;

UPDATE person
SET bday = '2000-Jan-01'
WHERE person_name = 'Lucy';

UPDATE person
SET bday = '2000-Feb-01'
WHERE person_name = 'Bob';

-- compare the results for the following queries:
select * from person
order by bday::date;

select * from person
order by bday;

Performance Tuning

There are ways to optimize performance of queries by reducing the amount of data being processed or by using SQL keywords that are optimized for certain processes.  Here are a few examples:

select name, age from person;

select * from person;

select * from 

(select employee_id, country_id from employee where join_date > '2020-01-01') t1

inner join

(employee_id, salary from compensation where comp_date > '2020-01-01') t2

on t1.employee_id = t2.employee_id;

Step 6 - Delete Data

The delete clause is used to delete data from a table.  Be careful when using the delete clause.  If you omit the 'where' clause, you'll end up deleting all data from the table.

-- delete row from country table where the ID of the country = 3

delete from country where id = 3;

Join Types

There are 3 types of joins that are available in SQL.  The following is a good diagram that explains the results from each join type:  The diagram and more information can be found here: https://www.w3schools.com/sql/sql_join.asp

Image Removed

--modify some data to get some results
UPDATE person
SET country_id = 4  --note that this country ID doesn't exist in the country table.
WHERE person_name = 'Bob';

-- modify some data to get the demo results
INSERT INTO country (id, country_name) VALUES (5, 'Mexico');

select * from 

employee t1

inner join 

compensation t2

on 

t1.

employee_id = 

t2.

Left Join

select * from person left join country on person.country_id = country.id;

employee_id

where join_date > '2020-01-01'

and comp_date > '2020-01-01';

Inserting, Updating, and Deleting Tables and Data

Create tables

--now doCreate athe rightfollowing join
selecttwo *tables fromfor person right join country on person.country_id = country.id;

Full Outer Join

select * from person full outer join country on person.country_id = country.id;

Sub-queries

this demo:
CREATE TABLE person ( id INTEGER PRIMARY KEY, person_name VARCHAR(100), income NUMERIC(7,2) );
CREATE TABLE country ( id INTEGER PRIMARY KEY, country_name VARCHAR(100) );

For more information on creating tables: https://www.postgresql.org/docs/9.2/sql-createtable.html

Insert Data into the Tables

select-- *Insert fromdata person
fullinto outer join (
	select * from country
	where id = 2) country_table
on person.country_id = country_table.id;

Case Statement

SELECTthe 'person' table:
INSERT INTO person (id, person_name, CASE 
WHEN income) BETWEENVALUES 50000 and 61000 THEN 1 
WHEN income BETWEEN 62000 and 71000 THEN 2 
WHEN income BETWEEN 72000 and 81000 THEN 3 END as income_category
FROM person;

Clean Up

(1, 'Sally', 60000);
INSERT INTO person (id, person_name, income) VALUES (2, 'Bob', 70000);
INSERT INTO person (id, person_name, income) VALUES (3, 'Lucy', 80000);

dropINSERT table person;
drop table country;

Performance Tuning

There are ways to optimize performance of queries by reducing the amount of data being processed or by using SQL keywords that are optimized for certain processes.  Here are a few examples:

INTO country (id, country_name) VALUES (1, 'Canada'), (2, 'USA'), (3, 'Mexico');

Update

Update the person table to have a column for the person's country:

Select particular columns to speed up performance.

ALTER TABLE person ADD column country_id integer;

Update the data:

select * from employee;

UPDATE person
SET country_id = 2
WHERE person_name = 'Lucy';

UPDATE 

person
SET country_id 

= 

1
WHERE person_name in ('Sally', 'Bob');

Delete Data

The delete clause is used to delete data from a table.  Be careful when using the delete clause.  If you omit the 'where' clause, you'll end up deleting all data from the table.

-- delete row from country table where the ID of the country = 3

delete from country where id = 3;

Drop tables

We can drop tables from our database by performing the following command.

drop table person;
drop table country;

Additional online resources:

Past Recordings:

SQL Introductory workshop conducted at KCNI Oct 2020: 

• Link: https://camh.webex.com/camh/ldr.php?RCID=a9ab57ca3058aab99d2d2d8de3361bab
• pw: KCNIsql1021

Contacts:

• adeel.ansari@camh.ca
• Natalia.Potapova@camh.ca
• Rachelle.Farkas@camh.ca