SQL Essentials

Mark McIlroy

www.blueskytechnology.com.au

www.markmcilroy.com

© Mark McIlroy 2016. All rights reserved.

ISBN 978-1492345831
Contents
1. Prerequisites..........................................................................................................................3
2. Instructions............................................................................................................................3
3. Joining tables.........................................................................................................................5
4. SQL........................................................................................................................................7
5. Single-table operations..........................................................................................................8
5.1 Selecting columns..............................................................................................................8
5.2 Selecting rows..................................................................................................................11
6. Sorting result tables.............................................................................................................13
7. Counting rows.....................................................................................................................14
8. Summing totals....................................................................................................................15
9. Cartesian Joins.....................................................................................................................19
10. Retrieving data from multiple tables...................................................................................20
11. Distinct values.....................................................................................................................24
12. Union...................................................................................................................................26
13. Subqueries...........................................................................................................................27
14. Updating data.......................................................................................................................28
14.1 Inserting records...........................................................................................................28
14.2 Updating records..........................................................................................................28
14.3 Deleting records...........................................................................................................29
15. NULL values.......................................................................................................................30
16. About the Author.................................................................................................................31
17. Appendix A – Implementation variations...........................................................................32
18. Appendix B – Summary of operators..................................................................................33
19. Appendix C – Other statements...........................................................................................34
20. Appendix D – Test environment.........................................................................................35
1. Prerequisites
This book assumes a familiarity with relational data tables.

Readers should have access to a query environment that allows viewing data tables and
running SQL queries against a database or data warehouse.

An online test environment is provided for readers of this book.

2. Instructions
A test environment is located at

www.markmcilroy.com/test_env/sql_test.php

You can try out the examples from the book in this environment.

The order of the notes is not significant.

Each of the queries listed in the notes is executable in the test environment.

Relational data tables

Relational data is stored in tables which can be represented in tabular format.

Customer_ID First_Name Surname Date_Of_Birth Postcode

0000001 Stephen Adjei 5/06/1988 4235
0000002 Sammy Adams 26/12/1983 5432
0000003 Linda Larigue 21/04/1976 2342
0000004 Sina Siva 24/07/1983 4342
0000005 Vangie Robinson 5/12/1988 5432
0000006 Christiana Majola 3/01/1973 2345
0000007 Olya Ayinoko 22/07/1968 9464
0000008 David Kellyn 14/08/1982 1242
0000009 Olusegun Aby 11/09/1982 4344
0000010 Maulesh Amoah 20/01/1965 5342
0000011 Raj Lucas 1/01/1973 6543
0000012 Emmanuel Crenshaw 6/04/1987 3456
0000013 Jessica Adesina 3/05/1977 2356
0000014 Selina Amoah 16/01/1982 5423
0000015 Ademola karmakar 11/03/1971 3474

4
Databases are typically used to store information such as customer records, transactions,
accounting records and so on.

Data processing is used extensively in the government and corporate business sectors.

5
3. Joining tables
Processing typically requires that information be retrieved from several different tables.

This is done using a process known as a ‘relational join’.

In the following example, details are extracted from the customer table and the transaction
table.

Transaction Table

Transaction_ID Transaction_Date Customer_ID Transaction_Code Product_Code Items Amount

0000001
0000001 6/01/2008 PURCHASE PCDE12 1 723.12
0000002 12/01/2008 0000001 PURCHASE PCDE27 1 324.12
0000003 4/01/2008 0000002 PURCHASE PCDE12 3 623.23
0000004 12/01/2008 0000002 CANCELLATION PCDE27 1 123.34
0000005 12/01/2008 0000002 PURCHASE PCDE12 2 423.23
0000006 5/01/2008 0000003 PURCHASE PCDE12 2 153.24
0000007 12/01/2008 0000004 REFUND PCDE43 1 233.22
0000008 21/01/2008 0000004 PURCHASE PCDE43 1 823.11

Customer Table

Customer_ID First_Name Surname Date_Of_Birth

1 Stephen Adjei 5/06/1988
2 Sammy Adams 26/12/1983
3 Linda Larigue 21/04/1976
4 Sina Siva 24/07/1983

6
Extracted Data

Trans_ID Trans_Date Cust_ID First_Name Surname Transaction_Code Prod_Code Items Amount

0000001 Stephen Adjei

0000001 6/01/2008 PURCHASE PCDE12 1 723.12
0000002 12/01/2008 0000001 Stephen Adjei PURCHASE PCDE27 1 324.12
0000003 4/01/2008 0000002 Sammy Adams PURCHASE PCDE12 3 623.23
0000004 12/01/2008 0000002 Sammy Adams CANCELLATION PCDE27 1 123.34
0000005 12/01/2008 0000002 Sammy Adams PURCHASE PCDE12 2 423.23
0000006 5/01/2008 0000003 Linda Larigue PURCHASE PCDE12 2 153.24
0000007 12/01/2008 0000004 Sina Siva REFUND PCDE43 1 233.22
0000008 21/01/2008 0000004 Sina Siva PURCHASE PCDE43 1 823.11

Note that data from one table has been duplicated in the columns of the result table.

In this example, the surname and first-name of the customer appears beside each transaction
relevant to that customer.

This is known as a ‘one-to-many’ join.

7
4. SQL
SQL, Structured Query Language, is a database query language that provides functions for
sorting, filtering and totaling information that is stored in relational data bases.

SQL is the basis for most query operations against large-scale data storage systems.

8
5. Single-table operations

Viewing a table can be done using a statement such as

select * from customer;

In this case the ‘*’ represents that all columns should be included in the result table.

5.1 Selecting columns

In many cases only a selection of columns is needed in the result set.

This is specified as follows.

select
surname,
first_name,
date_of_birth
from
customer;

Result

surname first_name date_of_birth

Adjei Stephen 1988-06-05
Adams Sammy 1983-12-26
Larigue Linda 1976-04-21
Patel Sabina 1973-07-24
Robinson Vangie 1988-12-05
Majola Christiana 1973-01-03
Ayinoko Olya 1968-07-22
Kellyn David 1982-08-14
Aby Olusegun 1982-09-11

9
Amoah Maulesh 1965-01-20
Lucas Raj 1973-01-01
Crenshaw Emmanuel 1987-04-06
Adesina Jessica 1977-05-03
Amoah Selina 1982-01-26
Karmakar Ademola 1971-03-11

The ‘from’ clause specifies the table that is used as the source of the data.

The ‘select’ column names indicate which columns to select.

Column definitions

In general the ‘select’ list will include a list of column names.

SQL also supports expressions in place of column names

For example

select
surname || ‘ ‘ || firstname as full_name,
from
customer

select
amount * items as value
from
transactions

Expressions can include the concatenation operator || which combines two text values and
mathematical operators.

Note: the || operator is the standard SQL operator for string concatenation.

Due to the variation of SQL used in the test environment, the function ‘concat()’ must be
used instead.

select
concat(first_name , ' ', surname) as full_name,

10
 from
customer

Result

Query Results

full_name
Stephen Adjei
Sammy Adams
Linda Larigue
Sabina Patel
Vangie Robinson
Christiana Majola
Olya Ayinoko
David Kellyn
Olusegun Aby
Maulesh Amoah
Christopher Lucas
Emmanuel Crenshaw
Jessica Adesina
Selina Amoah
Ademola Karmakar

11
5.2 Selecting rows

Generally rows are filtered so that only rows matching certain criteria are included in the
result set or the table calculations.

For example,

select
*
from
transactions
where
trans_date > '2008-01-04' and
product_code = 'PCDE12'

Result

trans_id trans_date cust_id transaction_code product_code items amount

1 2008-01-06 00:00:00 1 PURCHASE PCDE12 1 723.12
5 2008-01-12 00:00:00 2 PURCHASE PCDE12 2 423.23
6 2008-01-12 00:00:00 2 PURCHASE PCDE12 2 423.23
7 2008-01-05 00:00:00 3 PURCHASE PCDE12 2 153.24

Multiple conditions can be combined using ‘and’ and ‘or’.

Brackets should be used when the ‘where’ expression includes a combination of ‘and’ and
‘or’ expressions.

‘in’ operator

Comparisons can also be done with another table using the ‘in’ operator.

select
*
from
transactions
where
product_code in (select product_code from sample_list)

12
Result

Trans_id trans_date Cust_id transaction_code product_code items amount

1 2008-01-06 00:00:001 PURCHASE PCDE12 1 723.12
3 2008-01-04 00:00:002 PURCHASE PCDE12 3 623.23
5 2008-01-12 00:00:002 PURCHASE PCDE12 2 423.23
6 2008-01-12 00:00:002 PURCHASE PCDE12 2 423.23
7 2008-01-05 00:00:003 PURCHASE PCDE12 2 153.24
8 2008-01-12 00:00:004 REFUND PCDE43 1 233.22
9 2008-01-21 00:00:004 PURCHASE PCDE43 1 823.11

‘like’ operator

The ‘like’ operator allows for a selection of records matching similar patterns.

For example

select
*
from
customer
where
surname like ‘A%’

This selects all customers with surnames starting with ‘A’

Result

customer_id first_name surname date_of_birth postcode

1 Stephen Adjei 1988-06-05 4235
2 Sammy Adams 1983-12-26 5432
7 Olya Ayinoko 1968-07-22 9464
9 Olusegun Aby 1982-09-11 4344
10 Maulesh Amoah 1965-01-20 5342
13 Jessica Adesina 1977-05-03 2356
14 Selina Amoah 1982-01-26 5423

13
6. Sorting result tables
Rows returned from an SQL query may be returned in a random order.

The ordering of the rows can be specified using an ‘order by’ clause

For example

select
*
from
transactions
order by
customer_id,
trans_date desc;

Rows can be sorted in descending order by adding ‘desc’ after the column name.

Result

transa
customer
ction_i trans_date transaction_code product_code items amount
_id
d
2 2008-01-12 00:00:00 1 PURCHASE PCDE27 1 324.12
1 2008-01-06 00:00:00 1 PURCHASE PCDE12 1 723.12
4 2008-01-12 00:00:00 2 CANCELLATION PCDE27 1 425.54
5 2008-01-12 00:00:00 2 PURCHASE PCDE12 2 423.23
6 2008-01-12 00:00:00 2 PURCHASE PCDE12 2 423.23
3 2008-01-04 00:00:00 2 PURCHASE PCDE12 3 623.23
7 2008-01-05 00:00:00 3 PURCHASE PCDE12 2 153.24
9 2008-01-21 00:00:00 4 PURCHASE PCDE43 1 823.11
8 2008-01-12 00:00:00 4 REFUND PCDE43 1 233.22

14
7. Counting rows
count(*) can be used to count the number of records in a table.

For example

select
count(*)
from
transactions

Result

count(*)
9

A “where’ clause can be used to count a sub-set of the rows in the table

select
count(*)
from
transactions
where
trans_date > '2008-01-06'

Result

count(*)
6

Counting sets of records can be done using the ‘Group By’ clause.

15
8. Summing totals
A ‘group by’ clause can be used to calculate totals, averages and counts of records.

‘Group by’ is a complex use of SQL functionality and is not recommending for initial use.

Example

select
customer_id,
trans_date,
sum(amount)
from
transactions
group by
customer_id;

Result

customer_id trans_date sum(amount)

1 2008-01-06 00:00:00 1047.24
2 2008-01-04 00:00:00 1895.23
3 2008-01-05 00:00:00 153.24
4 2008-01-12 00:00:00 1056.33

SQL has an open syntax that will allow many combinations of queries to be written.

Only some combinations of clauses will produce meaningful results.

The following rules should be used to produce meaningful result sets when ‘group by’ is
used.

1. Select the ‘group by’ columns.

One row will be produced in the result set for each combination of the ‘group by’ columns.

For example,

16
 group by customer_id

Will produce one set of values for each customer.

group by customer_id, trans_date

Would produce a row for each date on which a customer transaction occurred.

2. Include each ‘group by’ column as a column in the ‘select’ section.

Example

select
customer_id
from
customer
group by
customer_id;

3. Select the aggregate functions

The aggregate functions include:

sum(column) Sum the column values

min(column) Select the minimum value
max(column) Select the maximum value
avg(column) Calculate the average value

count(*) Count the number of rows

count(column) Count the number of rows with non-NULL values
…

For example

select
customer_id,
count(*) as transaction_count,
sum(amount) as transaction_total
from
transactions
group by
customer_id;

17
Result

customer_id transaction_count transaction_total

1 2 1047.24
2 4 1895.23
3 1 153.24
4 2 1056.33

This query will return one row for each customer who has transaction records, with the
following columns:

customer_id The customer number

count(*) The number of transactions
sum(amount) The total value of the transactions

When an expression is used in place of a column name, the naming of the result column is
database-dependant.

In these cases it is preferable to name the result column.

Columns can also be renamed in the result set in this way.

4. Do not include additional columns in the ‘select’ column list

This may result in an undefined result set.

Aggregate functions in row selection

A ‘having’ clause can be added to a ‘group by’ clause when aggregate functions are used.

For example

select
customer_id,
count(*) as trans_count,
sum(amount) as trans_total
from
transactions
group by

18
 customer_id
having
sum(amount) > 200

Result

customer_id trans_count trans_total

1 2 1047.24
2 4 1895.23
4 2 1056.33

19
9. Cartesian Joins
A Cartesian join involves creating a result set containing all combinations of the records
from the input tables.

This is usually unintended.

For example

select
*
from
customer,
transaction

This would not be a meaningful result set, as transaction data would appear beside customer
details of a customer unrelated to the transaction.

The lack of a ‘where’ or ‘join’ clause will result in all combinations of records being
returned.

In cases where a Cartesian join is required, the number of records returned is m * n * p *…

Where m, n, p,… is the number of rows in the input tables.

20
10. Retrieving data from multiple tables
Most queries involve retrieving data from several input tables.

Tables must be connected using key fields.

These are generally columns such as customer number, product code, transaction date, etc.

Key fields identify a record, rather than being stored data such as amounts, text values, etc.

Joins may be specified in one of two ways.

Join syntax

select
t.customer_id,
t.trans_date,
c.postcode
from
transactions t
inner join
customer c
on
t.customer_id = c.customer_id;

Result

customer_id trans_date postcode

1 2008-01-06 00:00:00 4235
1 2008-01-12 00:00:00 4235
2 2008-01-04 00:00:00 5432
2 2008-01-12 00:00:00 5432
2 2008-01-12 00:00:00 5432
2 2008-01-12 00:00:00 5432
3 2008-01-05 00:00:00 2342
4 2008-01-12 00:00:00 4342
4 2008-01-21 00:00:00 4342

21
The join types are

Inner join Only records with matching keys are returned

Left join All records are returned from the first table, and matching records
from the second table

Right join All records are returned from the second table, and matching records
from the first table

Alias names

Alias names do not affect the result of a query however they can be useful in expressing the
query more simply.

For example

This query uses alias names ‘t’ and ‘c’

select
t.customer_id,
t.trans_date,
c.postcode
from
transactions as t
inner join
customer as c
on
t.customer_id = c.customer_id;

Alias names are necessary in the rare case in which an input table appears more than once in
a ‘select’ statement.

Also, if a column name appears in more than one input table, then an alias name should be
used to identify the relevant input table.

This problem typically results in an ‘ambiguous column name’ error.

Columns can also be specified using alias names in the format ‘a.*’.

This indicates that all columns from table ‘a’ should be included in the result set.

More than two join tables.

22
The following layout is recommended when more than two input tables are included in a join

select
t.customer_id,
t.trans_date,
c.postcode,
p.product_code as product,
cu.description as currency

from
transactions as t
inner join
customer as c
on
t.customer_id = c.customer_id
inner join
product as p
on
t.product_code = p.product_code
left join
currency as cu
on
cu.code = p.currency_code;

Result

customer_id trans_date postcode product currency

1 2008-01-06 00:00:00 4235 PCDE12 US Dollars
1 2008-01-12 00:00:00 4235 PCDE27 Hong Kong Dollars
2 2008-01-04 00:00:00 5432 PCDE12 US Dollars
2 2008-01-12 00:00:00 5432 PCDE27 Hong Kong Dollars
2 2008-01-12 00:00:00 5432 PCDE12 US Dollars
2 2008-01-12 00:00:00 5432 PCDE12 US Dollars
3 2008-01-05 00:00:00 2342 PCDE12 US Dollars
4 2008-01-12 00:00:00 4342 PCDE43 Japanese Yen
4 2008-01-21 00:00:00 4342 PCDE43 Japanese Yen

23
In the case of left joins and right joins, the order of tables in the query may affect the result
set. Each table is joined to the result of the previous joins. Field names in ‘on’ expressions
should only refer to tables that are specified earlier in the join list.

Where syntax

Joins can also be specified by listing multiple tables in the ‘from’ clause, and matching the
keys within the ‘where’ clause.

This syntax is equivalent to using ‘inner join’ on all the joined tables.

A ‘where’ format does not facilitate ‘left’ or ‘right’ joins

select
t.customer_id,
t.trans_date,
c.postcode,
p.product_code as product,
cu.description as currency

from
transactions as t,
customer as c,
product as p,
currency as cu
where
t.customer_id = c.customer_id and
cu.code = p.currency_code and
t.product_code = p.product_code

Multiple join fields

In some cases records will be identified by a single value such as transaction_id,

customer_number etc.

In other cases tables may be joined by a number or fields, such as product_class,

product_subclass

In these cases use a syntax similar to the following

24
 …
on
a.product_class = b.product_class and
a.product_subclass = b.product_subclass

11. Distinct values

Distinct values can be returned using the ‘distinct’ keyword or a ‘group by’ clause.

For example

select distinct
customer_id
from
transactions;

This query will return a list of the customer_id values that appear in the transaction table

Result

customer_id
1
2
3
4

This function can be used with any query, but is most useful when there is a single result
column or a small number of result columns.

If a count of these values is required, the ‘group by’ syntax should be used

select
customer_id,
count(*)
from
transactions
group by
customer_id;

25
Result

customer_id count(*)
1 2
2 4
3 1
4 2

26
12. Union

The ‘union’ statement can be used to combine the results of two queries into a single result
set.

select * from

(select * from sample_list

union all

select * from sample_list_ext) a

‘union all’ combines the two result sets, while ‘union’ selects only the distinct records

Result

id product_code
1 PCDE43
2 PCDE52
3 PCDE12
1 PCDE27
2 PCDE12

27
13. Subqueries
An SQL query can be used in place of a table name.

The query should be placed within brackets, and used in place of a table name within another
query.

For example

select
count(*)
from
transactions as t
inner join
(
select distinct
product_code
from
product
) as p
on
t.product_code = p.product_code

In this example a bracketed query has been used in place of a table name.

In this case, a count of records is calculated from customer records joined to product codes.

The statement within the brackets is equivalent to a table containing the same data.

28
14. Updating data

The following sections describe SQL statements for updating data.

In many cases it is not possible to recover data that is accidently altered or deleted.

Caution should be used when using these statements. For example

delete from transactions

Will delete all records from the database table ‘transactions’.

14.1 Inserting records

Individual rows can be inserted into a table using the following syntax

insert into
currexchange (name, amount, exchdate)
values
('name1', 12.52, '2003-02-01')

Importing large quantities of records is dependant on the functions provided by the database
environment.

14.2 Updating records

Tables can be updated using the following syntax.

update
currexchange
set
amount = 32.23
where
exchdate = ‘2003-02-01’

29
Implementations vary in their ability to perform updates on views created by joining several
tables.

14.3 Deleting records

Deletion takes the format

delete from table [where condition]

For example

delete from currexchange where exchdate < '2003-03-04'

30
15. NULL values
NULL values represent missing data.

This may indicate that a data item is not known, or is not relevant in that particular case.

Visual tools may display this result in several formats including NULL, (null), a blank field
etc.

In ‘where’ expressions the following syntax should be used

select
customer_id,
amount
from
transactions
where
amount is not NULL

Result

customer_id amount
1 723.12
1 324.12
2 623.23
2 425.54
2 423.23
2 423.23
3 153.24
4 233.22
4 823.11

31
16. About the Author

Mark McIlroy has an undergraduate degree in Computer Science and Applied Mathematics
from Monash University.

He has extensive experience consulting in the banking and government sectors in Australia
in large SQL data warehouse environments.

Mark lives with his wife in Melbourne, Australia.

32
17. Appendix A – Implementation variations
The SQL statements described here should be executable in most SQL environments.

Some differences may occur with issues such as specifying date constants.

For example

‘1990-04-12’

to_date( ‘01JUL2009’ )

etc.

Major implementations frequently have added syntax which is not compatible across
alternative implementations.

Examples include variations on join types such as OUTER JOIN, CROSS JOIN etc.

33
18. Appendix B – Summary of operators
Operators

Mathematical

* Multiplication
/ Division
+ Addition
- Subtraction

Relational

< Less than

<= Less than or equal to
> Greater than
>= Greater than or equal to
= Equal
<> Not equal
!= Not equal

String

|| Concatenate

Aggregate

sum Sum
avg Average
count Count
min Minimum value
max Maximum value
stdev Standard Deviation

34
19. Appendix C – Other statements

Other Issues

SQL includes the following groups of statements.

These statements are not widely used as these functions are more easily performed using
database administration tools.

Data Definition statements

Statements for creating tables and altering table formats

CREATE TABLE transactions ( id INTEGER NOT NULL,

transact_date DATE,
amount DOUBLE,
description VARCHAR(255),
PRIMARY KEY (‘id’) )

Administration statements

Statements for creating user accounts and assigning security privileges.

GRANT SELECT ON TABLE1 TO USERNAME1

Descriptive statements

Statements for returning the information about the database, such as the list of tables.

SHOW TABLES

35
20. Appendix D – Test environment

A test environment is located at

www.markmcilroy.com/test_env/sql_test.php

You can try out the examples from the book in this environment.

Tables

currency id, code, description

currexchange name, amount, exchdate

customer customer_id, first_name, surname, date_of_birth,

postcode

product id, product_code, currency_code, description

sample_list id, product_code

sample_list_ext id, product_code

transactions transaction_id, trans_date, customer_id,

transaction_code, product_code, items, amount

36