About the author

Rahul Batra was first introduced to programming in 1996 in GWBASIC, but he did not seriously foray into it till 2001 when he started learning C++. Along the way, there were dabblings in many other languages like C, Ruby, Perl and Java. He has worked on Oracle, MySQL, Sybase ASA, Ingres and SQLite.

Rahul has been programming professionally since 2006 and currently lives and works in Gurgaon, India.

",{"_345":346,"_42":43,"_46":-5,"_44":45,"_63":64,"_347":348,"_421":343},"previewContent","

\n Preface\n
\n About the author\n
\n Acknowledgements\n
\n 1. An Introduction to SQL\n
- \n 1.1 SQL Commands Classification\n
- \n 1.2 Explaining Tables\n
\n
\n 2. Getting your database ready\n
- \n 2.1 Using Ingres\n
- \n 2.2 Using SQLite\n
- \n 2.3 Creating your own database\n
- \n 2.4 Table Creation\n
- \n 2.5 Inserting data\n
- \n 2.6 Writing your first query\n
\n
\n 3. Constraints\n
- \n 3.1 Selective fields INSERT\n
- \n 3.2 Primary Key Constraint\n
- \n 3.3 Unique Key Constraint\n
- \n 3.4 Differences between a Primary Key and a Unique Key\n
\n
\n 4. Operations on Tables\n
- \n 4.1 Dropping Tables\n
- \n 4.2 Creating new tables from existing tables\n
- \n 4.3 Modifying tables\n
- \n 4.4 Verifying the result in Ingres\n
- \n 4.5 Verifying the result in other DBMS’s\n
- \n 4.6 Showing table information in SQLite\n
\n
\n 5. Writing Basic Queries\n
- \n 5.1 Selecting a limited number of columns\n
- \n 5.2 Ordering the results\n
- \n 5.3 Ordering using field abbreviations\n
- \n 5.4 Putting conditions with WHERE\n
- \n 5.5 Combining conditions\n
\n
\n 6. Manipulating Data\n
- \n 6.1 Inserting NULL’s\n
- \n 6.2 Inserting data into a table from another table\n
- \n 6.3 Updating existing data\n
- \n 6.4 Deleting data from tables\n
\n
\n 7. Organizing your data\n
- \n 7.1 Normalization\n
- \n 7.2 Atomicity\n
- \n 7.3 Repeating Groups\n
- \n 7.4 Splitting the table\n
\n
\n 8. Doing more with queries\n
- \n 8.1 Counting the records in a table\n
- \n 8.2 Column Aliases\n
- \n 8.3 Order of execution of SELECT queries\n
- \n 8.4 Using the LIKE operator\n
\n
\n 9. Calculated Fields\n
- \n 9.1 Mathematical calculations\n
- \n 9.2 String operations\n
- \n 9.3 Literal Values\n
\n
\n 10. Aggregation and Grouping\n
- \n 10.1 Aggregate Functions\n
- \n 10.2 Using DISTINCT with COUNT\n
- \n 10.3 Using MIN to find minimum values\n
- \n 10.4 Grouping Data\n
- \n 10.5 The HAVING Clause\n
\n
\n 11. Understanding Joins\n
- \n 11.1 What is a Join?\n
- \n 11.2 Alternative Join Syntax\n
- \n 11.3 Resolving ambiguity in join columns\n
- \n 11.4 Cross Joins\n
- \n 11.5 Self Joins\n
\n
\n 12. Subqueries\n
- \n 12.1 Types of subqueries\n
- \n 12.2 Using subqueries in INSERT statements\n
\n
\n 13. Working with Sets\n
- \n 13.1 Union\n
- \n 13.2 Intersection\n
- \n 13.3 Difference\n
\n
\n Further Reading\n
\n Appendix: Major Database Management Systems\n
\n Glossary\n

Preface

\n\n

Welcome to the third edition of A Primer on SQL. This edition features a new chapter on subqueries and a few other changes\nsuggested by readers. There is also some further information about using this text with SQLite which has continued to soar new\nheights in popularity, while Ingres has seen some slowness in recent times. However, this text remains database implementation\nagnostic.

\n\n

I hope that old and new readers find this text even more useful now in its presentation. I have tried to keep the spirit of the original text, a short introduction to the basics. As always, your questions, comments, criticism, encouragement and corrections are most welcome and you can e-mail me at rhlbatra[aht]hotmail[dot]com.

\n\n

Rahul Batra (25th February 2015)

\n\n

Preface to the first edition

\n\n

Welcome to the first edition of A Primer on SQL. As you would be able to see, the book is fairly short and is intended as an introduction to the basics of SQL. No prior experience with SQL is necessary, but some knowledge of working with computers in general is required. My purpose of writing this was to provide a gentle tutorial on the syntax of SQL, so that the reader is able to recognize the parts of queries they encounter and even be able to write simple SQL statements and queries themselves. The book however is not intended as a reference work or for a full time database administrator since it does not have an exaustive topic coverage.

\n\n

Your questions, comments, criticism, encouragement and corrections are most welcome and you can e-mail me at rhlbatra[aht]hotmail[dot]com. I’ll try answering all on-topic mails and will try to include suggestions, errors and omissions in future editions.

\n\n

Rahul Batra (8th October 2012)

\n\n\n

About the author

\n\n

Rahul has been programming professionally since 2006 and currently lives and works in Gurgaon, India.

\n\n\n

Acknowledgements

\n\n

This work would not have been completed without the support of my family and friends. First and foremost, I owe this book to my son.\nHe has given my life new meaning and direction. A thank you is in order for my wife Pria, who not only acted as an editor but also\nconstantly supported and cheered me on to complete it. Many thanks to my parents too, who got me a computer early in life to start\ntinkering around with and for constantly encouraging me to pursue my dreams.

\n\n

Thanks also go out to my sister, niece and nephew (may you have beautiful lives ahead) and my friends for bringing much happiness\ninto my life. Finally I would like to acknowledge the contribution of my teachers who helped me form my computing knowledge.

\n\n

I would also like to acknowledge the contributions of the following readers who suggested improvements and caught errors I had\nmissed: Keith Thompson, Nathan Adams, Paul Guilbault, Jim Noh and Sean Farrell.

\n\n\n\n

\n1. An Introduction to SQL

\n\n

A database is nothing but a collection of organized data. It doesn’t have to be in a digital format to be called a database. A telephone directory is a good example, which stores data about people and organizations with a contact number. Software which is used to manage a digital database is called a Database Management System (DBMS).

\n\n

The most prevalent database organizational model is the Relational Model, developed by Dr. E F Codd in his groundbreaking research paper - A Relational Model of Data for Large Shared Data Banks.In this model, data to be stored is organized as rows inside a table with the column headings specifying the corresponding type of data stored. This is not unlike a spreadsheet where the first row can be thought of as column headings and the subsequent rows storing the actual data.

\n\n\n\n

SQL stands for Structured Query Language and it is the de-facto standard for interacting with relational databases. Almost all database management systems you’ll come across will have a SQL implementation. SQL was standardized by the American National Standards Institute (ANSI) in 1986 and has undergone many revisions, most notably in 1992 and 1999. However, all DBMS’s do not strictly adhere to the standard defined but rather remove some features and add others to provide a unique feature set. Nonetheless, the standardization process has been helpful in giving a uniform direction to the vendors in terms of their database interaction language.

\n\n

\n1.1 SQL Commands Classification

\n\n

SQL is a language for interacting with databases. It consists of a number of commands with further options to allow you to carry out your operations with a database. While DBMS’s differ in the command subset they provide, usually you would find the classifications below.

\n\n

\nData Definition Language (DDL) : CREATE TABLE, ALTER TABLE, DROP TABLE etc.\n
These commands allow you to create or modify your database structure.
\n
\nData Manipulation Language (DML) : INSERT, UPDATE, DELETE\n
These commands are used to manipulate data stored inside your database.
\n
\nData Query Language (DQL) : SELECT\n
Used for querying or selecting a subset of data from a database.
\n
\nData Control Language (DCL) : GRANT, REVOKE etc.\n
Used for controlling access to data within a database, commonly used for granting user privileges.
\n
\nTransaction Control Commands : COMMIT, ROLLBACK etc.\n
Used for managing groups of statements as a unit of work.
\n

\n\n

Besides these, your database management system may give you other sets of commands to work more efficiently or to provide extra features. But it is safe to say that the ones above would be present in almost all DBMS’s you encounter.

\n\n

\n1.2 Explaining Tables

\n\n

A table in a relational database is nothing but a matrix of data where the columns describe the type of data and the row contains the actual data to be stored. Have a look at the figure below to get a sense of the visualization of a table in a database.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a table describing Programming Languages
id	language	author	year
1	Fortran	Backus	1955
2	Lisp	McCarthy	1958
3	Cobol	Hopper	1959

\n\n

The above table stores data about programming languages. It consists of 4 columns (id, language, author and year) and 3 rows. The formal term for a column in a database is a field and a row is known as a record.

\n\n

There are two things of note in the figure above. The first one is that, the id field effectively tells you nothing about the programming language by itself, other than its sequential position in the table. The second is that though we can understand the fields by looking at their names, we have not formally assigned a data type to them i.e. we have not restricted (not yet anyways) whether a field should contain alphabets or numbers or a combination of both.

\n\n

The id field here serves the purpose of a primary key in the table. It makes each record in the table unique and its advantages will become clearer in chapters to come. But for now consider this, what if a language creator made two languages in the same year; we would have a difficult time narrowing down on the records. An id field usually serves as a good primary key since it’s guaranteed to be unique, but usage of other fields for this purpose is not restricted.

\n\n

Just like programming languages, SQL also has data types to define the kind of data that will be stored in its fields. In the table given above, we can see that the fields language and author must store English language characters. Thus their data type during table creation should be specified as varchar which stands for variable number of characters.

\n\n

The other commonly used data types you will encounter in subsequent chapters are:

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Fixed length characters	char
Integer values	int
Decimal numbers	decimal
Date data type	date

\n\n\n

\n2. Getting your database ready

\n\n

\n2.1 Using Ingres

\n\n

The best way to learn SQL is to practice writing commands on a real relational database. In this book SQL is taught using a product called Ingres. The reasons for choosing Ingres are simple - it comes in a free and open source edition, it’s available on most major platforms and it’s a full-fledged enterprise class database with many features. However, any relational database product that you can get your hands on should serve you just fine. There might be minor incompatibilities between different vendors, so if you choose something else to practice on while reading this book, it would be a good idea to keep the database vendor’s user manual handy.

\n\n

Since this text deals largely with teaching SQL in a product independent manner, rather than the teaching of Ingres per se, details with respect to installation and specific operations of the product will be kept to a minimum. Emphasis is instead placed on a few specific steps that will help you to get working on Ingres as fast as possible.

\n\n

The current version of Ingres during the writing of the book was 10.1 and the Community Edition has been used on a Windows box for the chapters to follow. The installation itself is straightforward like any other Windows software. However if you are unsure on any option, ask your DBA (database administrator, in case one is available) or if you are practicing on a home box - select the ‘Traditional Ingres’ mode and install the Demo database when it asks you these questions. Feel free to refer to the Ingres installation guide that is available on the web at the following location. Ingres Installation Guide

\n\n

If your installation is successful, you should be able to start the Ingres Visual DBA from the Start Menu. This utility is a graphical user interface to manage your Ingres databases, but we will keep the usage of this to a minimum since our interest lies in learning SQL rather than database administration.

\n\n

\n2.2 Using SQLite

\n\n

If installing Ingres seems like a daunting task, you are in luck. There is a very credible, free alternative database for you to practice on. It is called SQLite and it’s creator D. Richard Hipp has generously licensed it in the public domain. You can download it from the SQLite Download page.

\n\n

If you are using Microsoft Windows, you are looking for the section titled Precompiled Binaries for Windows. Download the SQLite DLL zip archive, named like sqlite-dll-win32-x86-xxxxxxx.zip, which contains SQLite but not a way to interact with it. For that you must download the SQLite shell, named like sqlite-shell-win32-x86-xxxxxxx.zip, which will allow us to create and query SQLite databases through the command line.

\n\n

Extract both these archives into the same directory and you are done installing SQLite. Your folder should now contain atleast three files - sqlite3.dll, sqlite3.def, sqlite3.exe. The last one launches the command shell used to interact with SQLite databases.

\n\n

\n2.3 Creating your own database

\n\n

Most database management systems, including Ingres, allow you to create multiple databases. For practice purposes it’s advisable to create your own database, so that you are free to perform any operations on it.

\n\n

Most database systems differ in the way they provide database creation facilities. Ingres achieves the same by providing you multiple ways to do this, including through the Visual DBA utility. However for didactic purposes, we will instead use a command operation to create our database. Open up the Ingres Command Prompt from the program menu (usually found inside Start Menu->Programs->Ingres for Microsoft Windows systems), and enter the command as below.

\n\n

Listing: using createdb and its sample output

\n\n

1 C:\\\\Documents and Settings\\\\rahulb>createdb testdb\n2 Creating database 'testdb' . . .\n3   Creating DBMS System Catalogs . . .\n4   Modifying DBMS System Catalogs . . .\n5   Creating Standard Catalog Interface . . .\n6   Creating Front-end System Catalogs . . .\n7 Creation of database 'testdb' completed successfully.\n

\n\n\n\n

The command createdb is used to create a database which will serve as a holding envelope for your tables. In the example and output shown above, we created a database called testdb for our use. You (or more specifically your system login) are now the owner of this database and have full control of entities within it. This is analogous to creating a file in an operating system where the creator gets full access control rights and may choose to give other users and groups specific rights.

\n\n

If you are using SQLite, fire up the command shell and you will be greeted with a window with the text displayed below.

\n\n

1 SQLite version 3.8.2 2013-12-06 14:53:30\n2 Enter \".help\" for instructions\n3 Enter SQL statements terminated with a \";\"\n4 sqlite>\n

\n\n\n\n

Here we enter our .open command to both create a SQLite database or open it in case it already exists.

\n\n

1 .open testdb\n

\n\n\n\n

If you are using Linux, SQLite does not come with the .open command on it. Instead you directly write the database name on the\nterminal immediately after the interactive SQL shell program name like below.

\n\n

1 sqlite3 testdb\n

\n\n\n\n\n\n

\n2.4 Table Creation

\n\n

We have already explored the concept of a table in a relational model. It is now time to create one using a standard SQL command - CREATE TABLE.

\n\n\n\n

Listing: General Syntax of a CREATE TABLE statement

\n\n

1 CREATE TABLE <Table_Name>\n2 (<Field 1> <Data Type>,\n3  <Field 2> <Data Type>,\n4  \\. \\. \\.\n5  <Field N> <Data Type>);\n

\n\n\n\n

This is the simplest valid statement that will create a table for you, devoid of any extra options. We’ll further this with clauses and constraints as we go along, but for now let us use this general syntax to actually create the table of programming languages we introduced in Chapter 1.

\n\n

The easiest way to get started with writing SQL statements in Ingres is to use their Visual SQL application which gives you a graphical interface to write statements and view output. The usual place to find it on a Windows system is Start -> Programs -> Ingres -> Ingres II -> Other Utilities.

\n\n

When you open it up, it gives you a set of dropdown boxes on the top half of the window where you can select the database you wish to work upon and other such options. Since we’ll be using the same database we created previously (testdb), go ahead and select the options as specified below.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Default User	<your username>
Default Server	INGRES
Database	testdb

\n\n

The actual SQL statement you would be writing to create your table is given below.

\n\n

Listing: Creating the programming languages table

\n\n

1 CREATE TABLE proglang_tbl (\n2 id        INTEGER,\n3 language  VARCHAR(20),\n4 author    VARCHAR(25),\n5 year      INTEGER);\n

\n\n\n\n

Press the ‘Go’ or F5 button when you’re done entering the statement in full. If you get no errors back from Visual SQL, then congratulations are in order since you’ve just created your first table.

\n\n

The statement by itself is simple enough since it resembles the general syntax of CREATE TABLE we discussed beforehand. It is interesting to note the data types chosen for the fields. Both id and year are specified as integers for simplicity, even though there are better alternatives. The language field is given a space of 20 characters to store the name of the programming language while the author field can hold 25 characters for the creator’s name.

\n\n

The semicolon at the last position is the delimiter for SQL statements and it marks the end of a statement.

\n\n

The same CREATE TABLE statement also works fine for SQLite and is written in the SQLite command shell itself.

\n\n

\n2.5 Inserting data

\n\n

The table we have just created is empty so our task now becomes insertion of some sample data inside it. To populate this data in the form of rows we use the DML command INSERT, whose general syntax is given below.

\n\n

Listing: General syntax of INSERT TABLE

\n\n

1 INSERT INTO <Table Name>\n2 VALUES ('Value1', 'Value2', ...);\n

\n\n\n\n

Fitting some sample values into this general syntax is simple enough, provided we keep in mind the structure of the table we are trying to insert the row in. For populating the proglang_tbl with rows like we saw in chapter 1, we would have to use three INSERT statements as below.

\n\n

Listing: Inserting data into the proglang_tbl table

\n\n

1 INSERT INTO proglang_tbl VALUES (1, 'Fortran', 'Backus', 1955);\n2 INSERT INTO proglang_tbl VALUES (2, 'Lisp', 'McCarthy', 1958);\n3 INSERT INTO proglang_tbl VALUES (3, 'Cobol', 'Hopper', 1959);\n

\n\n\n\n

If you do not receive any errors from Ingres Visual SQL (or the SQL interface for your chosen DBMS), then you have managed to successfully insert 3 rows of data into your table. Notice how we’ve carefully kept the ordering of the fields in the same sequence as we used for creating our table. This strict ordering limitation can be removed and we will see how to achieve that in a little while.

\n\n

\n2.6 Writing your first query

\n\n

Let us now turn our attention to writing a simple query to check the results of our previous operations in which we created a table and inserted three rows of data into it. For this, we would use a Data Query Language (DQL) command called SELECT.

\n\n

A query is simply a SQL statement that allows you to retrieve a useful subset of data contained within your database. You might have noticed the INSERT and CREATE TABLE commands were referred to as statements, but a fetching operation with SELECT falls under the query category.

\n\n

Most of your day to day operations in a SQL environment would involve queries, since you’d be creating the database structure once (modifying it only on a need basis) and inserting rows only when new data is available. While a typical SELECT query is fairly complex with many clauses, we will begin our journey by writing down a query just to verify the contents of our table. The general syntax of a simple query is given below.

\n\n

Listing: General Syntax of a simple SQL query

\n\n

1 SELECT <Selection> FROM <Table Name>;\n

\n\n\n\n

Transforming this into our result verification query is a simple task. We already know the table we wish to query - proglang_tbl and for our selection we would use * (star), which will select all rows and fields from the table.

\n\n

1 SELECT * FROM proglang_tbl;\n

\n\n\n\n

The output of this query would be all the (3) rows displayed in a matrix format just as we intended. If you are running this through Visual SQL on Ingres, you would get a message at the bottom saying - Total Fetched Row(s): 3.

\n\n\n

\n3. Constraints

\n\n

A constraint is a rule that you apply or abide by while doing SQL operations. They are useful in cases where you wish to make the data inside your database more meaningful and/or structured. Consider the example of the programming languages table - every programming language that has been created, must have an author (whether a single person, or a couple or a committee). Similarly it should have a year when it was introduced, be it the year it first appeared as a research paper or the year a working compiler for it was written. In such cases, it makes sense to create your table in such a way that certain fields do not accept a NULL (empty) value.

\n\n

We now modify our previous CREATE TABLE statement so that we can apply the NULL constraint to some fields.

\n\n

Listing: Creating a table with NULL constraints

\n\n

1 CREATE TABLE proglang_tblcopy (\n2 id        INTEGER     NOT NULL,\n3 language  VARCHAR(20) NOT NULL,\n4 author    VARCHAR(25) NOT NULL,\n5 year      INTEGER     NOT NULL,\n6 standard  VARCHAR(10) NULL);\n

\n\n\n\n

We see in this case that we have achieved our objective of creating a table in which the field’s id, language, author and year cannot be empty for any row, but the new field standard can take empty values. We now go about trying to insert new rows into this table using an alternative INSERT syntax.

\n\n

\n3.1 Selective fields INSERT

\n\n

From our last encounter with the INSERT statement, we saw that we had to specify the data to be inserted in the same order as specified during the creation of the table in question. We now look at another variation which will allow us to overcome this limitation and handle inserting rows with embedded NULL values in their fields.

\n\n

Listing: General Syntax of INSERT with selected fields

\n\n

 1 INSERT INTO <Table_Name>\n 2 \n 3 (<Field Name 1>,\n 4  <Field Name 2>,\n 5  . . .\n 6  <Field Name N>)\n 7 \n 8 VALUES\n 9 \n10 (<Value For Field 1>,\n11  <Value For Field 2>,\n12  . . .\n13  <Value For Field N>);\n

\n\n\n\n

Since we specify the field order in the statement itself, we are free to reorder the values sequence in the same statement thus removing the first limitation. Also, if we wish to enter a empty (NULL) value in any of the fields for a record, it is easy to do so by simply not including the field’s name in the first part of the statement. The statement would run fine without specifying any fields you wish to omit provided they do not have a NOT NULL constraint attached to them. We now write some INSERT statements for the proglang_tblcopy table, in which we try to insert some languages which have not been standardized by any organizations and some which have been.

\n\n

Listing: Inserting new data into the proglang_tblcopy table

\n\n

1 INSERT INTO proglang_tblcopy (id, language, author, year, standard)\n2 VALUES (1, 'Prolog', 'Colmerauer', '1972', 'ISO');\n3 \n4 INSERT INTO proglang_tblcopy (id, language, author, year)\n5 VALUES (2, 'Perl', 'Wall', '1987');\n6 \n7 INSERT INTO proglang_tblcopy (id, year, standard, language, author)\n8 VALUES (3, '1964', 'ANSI', 'APL', 'Iverson');\n

\n\n\n\n

When you run this through your SQL interface, 3 new rows would be inserted into the table. Notice the ordering of the third row; it is not the same sequence we used to create the table. Also Perl has not been standardized by an international body, so we do not specify the field name itself while doing the INSERT operation.

\n\n

To verify the results of these statements and to make sure that the correct data went into the correct fields, we run a simple query as before.

\n\n

1 SELECT * FROM proglang_tblcopy;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Result of the query run on proglang_tblcopy
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI

\n\n

\n3.2 Primary Key Constraint

\n\n

A primary key is used to make each record unique in atleast one way by forcing a field to have unique values. They do not have to be restricted to only one field, a combination of them can also be defined as a primary key for a table. In our programming languages table, the id field is a good choice for applying the primary key constraint. We will now modify our CREATE TABLE statement to incorporate this.

\n\n

Listing: a CREATE TABLE statement with a primary key

\n\n

1 CREATE TABLE proglang_tbltmp (\n2 id        INTEGER     NOT NULL  PRIMARY KEY,\n3 language  VARCHAR(20) NOT NULL,\n4 author    VARCHAR(25) NOT NULL,\n5 year      INTEGER     NOT NULL,\n6 standard  VARCHAR(10) NULL);\n

\n\n\n\n

ID fields are usually chosen as primary fields. Note that in this particular table, the language field would have also worked, since a language name is unique. However, if we have a table which describes say people - since two people can have the same name, we usually try to find a unique field like their SSN number or employee ID number.

\n\n

\n3.3 Unique Key Constraint

\n\n

A unique key like a primary key is also used to make each record inside a table unique. Once you have defined the primary key of a table, any other fields you wish to make unique is done through this constraint. For example, in our database it now makes sense to have a unique key constraint on the language field. This would ensure none of the records would duplicate information about the same programming language.

\n\n

Listing: a CREATE TABLE statement with a primary key and a unique constraint

\n\n

1 CREATE TABLE proglang_tbluk (\n2 id        INTEGER     NOT NULL  PRIMARY KEY,\n3 language  VARCHAR(20) NOT NULL  UNIQUE,\n4 author    VARCHAR(25) NOT NULL,\n5 year      INTEGER     NOT NULL,\n6 standard  VARCHAR(10) NULL);\n

\n\n\n\n

Note that we write the word UNIQUE in front of the field and omit the KEY in this case. You can have as many fields with unique constraints as you wish.

\n\n

\n3.4 Differences between a Primary Key and a Unique Key

\n\n

You might have noticed that the two constraints discussed above are similar in their purpose. However, there are a couple of differences between them.

\n\n

A primary key field cannot take on a NULL value, whereas a field with a unique constraint can. However, there can be only one such record since each value must be unique due to the very definition of the constraint.
You are allowed to define only one primary key constraint but you can apply the unique constraint to as many fields as you like.

\n\n\n

\n4. Operations on Tables

\n\n

You might have noticed that we keep on making new tables whenever we are introducing a new concept. This has had the not-so desirable effect of populating our database with many similar tables. We will now go about deleting unneeded tables and modifying existing ones to suit our needs.

\n\n

\n4.1 Dropping Tables

\n\n

The deletion of tables in SQL is achieved through the DROP TABLE command. We will now drop any superfluous tables we have created during the previous lessons.

\n\n

Listing: dropping the temporary tables we created

\n\n

1 DROP TABLE proglang_tbl;\n2 \n3 DROP TABLE proglang_tblcopy;\n4 \n5 DROP TABLE proglang_tbltmp;\n

\n\n\n\n

\n4.2 Creating new tables from existing tables

\n\n

You might have noticed that we have dropped the proglang_tbl table and we now have with us only the proglang_tbluk table which has all the necessary constraints and fields. The latter’s name was chosen when we were discussing the unique key constraint, but it now seems logical to migrate this table structure (and any corresponding data) back to the name proglang_tbl. We achieve this by creating a copy of the table using a combination of both CREATE TABLE and SELECT commands and learn a new clause AS.

\n\n

Listing: general syntax for creating a new table from an existing one

\n\n

1 CREATE TABLE <New Table> AS SELECT <Selection> FROM <Old Table>;\n

\n\n\n\n

Since our proglang_tbluk contains no records, we will push some sample data in it so that we can later verify whether the records themselves got copied or not. Notice that we would have to give the field names explicitly, else the second row (which contains no standard field value) would give an error similar to ‘number of target columns must equal the number of specified values’ in Ingres.

\n\n

Listing: inserting some data into the proglang_tbluk table

\n\n

1 INSERT INTO proglang_tbluk (id, language, author, year, standard)\n2 VALUES (1, 'Prolog', 'Colmerauer', '1972', 'ISO');\n3 \n4 INSERT INTO proglang_tbluk (id, language, author, year)\n5 VALUES (2, 'Perl', 'Wall', '1987');\n6 \n7 INSERT INTO proglang_tbluk (id, year, standard, language, author)\n8 VALUES (3, '1964', 'ANSI', 'APL', 'Iverson');\n

\n\n\n\n

To create an exact copy of the existing table, we use the same selection criteria as we have seen before - * (star). This will select all the fields from the existing table and create the new table with them alongwith any records. It is possible to use only a subset of fields from the old table by modifying the selection criteria and we will see this later.

\n\n

Listing: recreating a new table from an existing one

\n\n

1 CREATE TABLE proglang_tbl AS SELECT * FROM proglang_tbluk;\n

\n\n\n\n

We now run a simple SELECT query to see whether our objective was achieved or not.

\n\n

1 SELECT * FROM proglang_tbl;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Result of the query run on proglang_tbl
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI

\n\n

\n4.3 Modifying tables

\n\n

After a table has been created, you can still modify its structure using the ALTER TABLE command. What we mean by modify is that you can change field types, sizes, even add or delete columns. There are some rules you have to abide by while altering a table, but for now we will see a simple example to modify the field author for the proglang_tbl table.

\n\n

Listing: General syntax of a simple ALTER TABLE command

\n\n

1 ALTER TABLE <Table name> <Operation> <Field with clauses>;\n

\n\n\n\n

We already know that we are going to operate on the proglang_tbl table and the field we wish to modify is author which should now hold 30 characters instead of 25. The operation to choose in this case is ALTER which would modify our existing field.

\n\n

Listing: Altering the author field

\n\n

1 ALTER TABLE proglang_tbl ALTER author varchar(30);\n

\n\n\n\n

\n4.4 Verifying the result in Ingres

\n\n

While one option to verify the result of our ALTER TABLE command is to run an INSERT statement with the author’s name greater than 25 characters and verify that we get no errors back, it is a tedious process. In Ingres specifically, we can look at the Ingres Visual DBA application to check the columns tab in the testdb database. However, another way to verify the same using a console tool is the isql command line tool available through the Ingres Command Prompt we used earlier for database creation.

\n\n

To launch isql (which stands for Interactive SQL) using the Ingres command prompt we type:

\n\n

1 isql testdb\n

\n\n\n\n

The first argument we write is the database we wish to connect to. The result of running this command is an interactive console window where you would be able to write SQL statements and verify the results much like Visual SQL. The difference between the two (other than the obvious differences in the user interface) is that isql allows you access to the HELP command, which is what we will be using to verify the result of our ALTER TABLE statement. In the interaction window that opens up, we write the HELP command as below and the subsequent box shows the output of the command.

\n\n

HELP TABLE proglang_tbl;

\n\n

Figure: the result of running the HELP TABLE command

\n\n

 1 Name:                 proglang_tbl\n 2 Owner:                rahulb\n 3 Created:              20-feb-2012 17:04:28     \n 4 Location:             ii_database\n 5 Type:                 user table\n 6 Version:              II10.0\n 7 Page size:            8192\n 8 Cache priority:       0\n 9 Alter table version:  4\n10 Alter table totwidth: 76\n11 Row width:            76\n12 Number of rows:       3\n13 Storage structure:    heap\n14 Compression:          none\n15 Duplicate Rows:       allowed\n16 Number of pages:      3\n17 Overflow data pages:  0\n18 Journaling:           enabled after the next checkpoint\n19 Base table for view:  no\n20 Permissions:          none\n21 Integrities:          none\n22 Optimizer statistics: none\n23 \n24 Column Information:\n25 \n26 |Column Name |Type    |Length |Nulls |Defaults |Key Seq |\n27 |------------|--------|-------|------|---------|--------|\n28 |id          |integer |4      |no    |no       |        |\n29 |language    |varchar |20     |no    |no       |        |\n30 |author      |varchar |30     |yes   |null     |        |\n31 |year        |integer |4      |no    |no       |        |\n32 |standard    |varchar |10     |yes   |null     |        |\n33 \n34 Secondary indexes:    none\n

\n\n\n\n

While there is a lot of information in the result, we are currently interested in the Column Information section which now displays the new length of the author field, i.e. 30. But it is also important to note that our ALTER TABLE statement just removed the not-null constraint from the field. To retain the same, we would have to specify the constraint in the alter command since the default behavior is to allow NULL values.

\n\n

\n4.5 Verifying the result in other DBMS’s

\n\n

The HELP command we just saw is specific to the Ingres RDBMS, it is not a part of the SQL standard. To achieve the same objective on a different RDBMS like Oracle, you are provided with the DESCRIBE command which allows you to view a table definition. While the information this command show may vary from one DBMS to another, they at least show the field name, its data type and whether or not NULL values are allowed for the particular field. The general synatax of the command is given below.

\n\n

Listing: the general syntax of the DESCRIBE statement

\n\n

1 DESCRIBE <table name>;\n

\n\n\n\n

\n4.6 Showing table information in SQLite

\n\n

SQLite as of the writing of this text does not support modification to column sizes in a table using ALTER TABLE. It does however allow you to view table and column information.

\n\n

SQLite has it’s own special dot syntax commands which allow certain useful database management tasks. We have already seen the .open command used to create and open a database. Similarly we can use the .schema command to get table information.

\n\n

Listing: showing table and column information in SQLite

\n\n

1 sqlite> .schema proglang_tbl\n2 \n3 CREATE TABLE proglang_tbl (\n4 id INTEGER NOT NULL PRIMARY KEY,\n5 language VARCHAR(20) NOT NULL UNIQUE,\n6 author VARCHAR(25) NOT NULL,\n7 year INTEGER NOT NULL,\n8 standard VARCHAR(10) NULL);\n

\n\n\n\n\n

\n5. Writing Basic Queries

\n\n

A query is a SQL statement that is used to extract a subset of data from your database and presents it in a readable format. As we have seen previously, the SELECT command is used to run queries in SQL. You can further add clauses to your query to get a filtered, more meaningful result. Since the majority of operations on a database involve queries, it is important to understand them in detail. While this chapter will only deal with queries run on a single table, you can run a SELECT operation on multiple tables in a single statement.

\n\n

\n5.1 Selecting a limited number of columns

\n\n

We have already seen how to extract all the data from a table when we were verifying our results in the previous chapters. But as you might have noted - a query can be used to extract a subset of data too. We first test this by limiting the number of fields to show in the query output by not specifying the * selection criteria, but by naming the fields explicitly.

\n\n

Listing: selecting a subset of fields from a table

\n\n

1 SELECT language, year FROM proglang_tbl;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of running the chosen fields SELECT query
language	year
Prolog	1972
Perl	1987
APL	1964

\n\n

You can see that the query we constructed mentioned the fields we wish to see, i.e. language and year. Also note that the result of this query is useful by itself as a report for looking at the chronology of programming language creation. While this is not a rule enforced by SQL or a relation database management system, it makes sense to construct your query in such a way that the meaning is self-evident if the output is meant to be read by a human. This is the reason we left out the field id in the query, since it has no inherent meaning to the reader except if they wish to know the sequential order of the storage of records in the table.

\n\n

\n5.2 Ordering the results

\n\n

You might have noticed that in our previous query output, the languages were printed out in the same order as we had inserted them. But what if we wanted to sort the results by the year the language was created in. The chronological order might make more sense if we wish to view the development of programming languages through the decades. In such cases, we take the help of the ORDER BY clause. To achieve our purpose, we modify our query with this additional clause.

\n\n

Listing: Usage of the ORDER BY clause

\n\n

1 SELECT language, year FROM proglang_tbl ORDER BY year;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of the ordered SELECT query
language	year
APL	1964
Prolog	1972
Perl	1987

\n\n

The astute reader will notice that the output of our ORDER BY clause was ascending. To reverse this, we add the argument DESC to our ORDER BY clause as below.

\n\n

Listing: Usage of the ORDER BY clause with the DESC argument

\n\n

1 SELECT language, year FROM proglang_tbl ORDER BY year DESC;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of the ordered SELECT query in descending order
language	year
Perl	1987
Prolog	1972
APL	1964

\n\n

\n5.3 Ordering using field abbreviations

\n\n

A useful shortcut in SQL involves ordering a query result using an integer abbreviation instead of the complete field name. The abbreviations are formed starting with 1 which is given to the first field specified in the query, 2 to the second field and so on. Rewriting our above query to sort the output by descending year, we get:

\n\n

1 SELECT language, year FROM proglang_tbl ORDER BY 2 DESC;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of the ordered SELECT query in descending order using field abbreviations
language	year
Perl	1987
Prolog	1972
APL	1964

\n\n

The 2 argument given to the ORDER BY clause signifies ordering by the second field specified in the query, namely year.

\n\n

\n5.4 Putting conditions with WHERE

\n\n

We have already seen how to select a subset of data available in a table by limiting the fields queried. We will now limit the number of records retrieved in a query using conditions. The WHERE clause is used to achieve this and it can be combined with explicit field selection or ordering clauses to provide meaningful output.

\n\n

For a query to run successfully, it must have atleast two parts - the SELECT and the FROM clause. After this we place the optional WHERE condition and then the ordering clause. Thus, if we wanted to see the programming language (and it’s author) which was standardized by ANSI, we’d write our query as below.

\n\n

Listing: Using a WHERE conditional

\n\n

1 SELECT language, author FROM proglang_tbl WHERE standard = 'ANSI';\n

\n\n\n\n

As you may have noticed, the query we forulated specified the language and author fields, but the condition was imposed on a separate field altogether - standard. Thus we can safely say that while we can choose what columns to display, our conditionals can work on a record with any of its fields.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of the SELECT query with a WHERE conditional clause
language	author
APL	Iverson

\n\n

You are by no means restricted to use = (equals) for your conditions. It is perfectly acceptable to choose other operators like < and >. You can also include the ORDER BY clause and sort your output. An example is given below.

\n\n

Listing: Combining the WHERE and ORDER BY

\n\n

1 SELECT language, author, year FROM proglang_tbl WHERE year > 1970 ORDER BY autho\\\n2 r;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Output of the SELECT query with a WHERE and ORDER BY
language	author	year
Prolog	Colmerauer	1972
Perl	Wall	1987

\n\n

Notice that the output only shows programming languages developed after 1970 (atleast according to our database). Also since the ordering is done by a varchar field, the sorting is done alphabetically in an ascending order.

\n\n

\n5.5 Combining conditions

\n\n

If we can only specify one condition using the WHERE clause, it will fulfill only a tiny fraction of real world requirements. We can however construct complex conditions using the boolean operators AND and OR.

\n\n

When we want our resultset to satisfy all of the multiple conditions, we use the AND operator.

\n\n

Listing: using the AND operator to combine conditions

\n\n

1 SELECT language, author, year FROM proglang_tbl\n2 WHERE year > 1970 AND standard IS NULL;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	author	year
Perl	Wall	1987
Tcl	Ousterhout	1988

\n\n

The result satisfies both the conditions we specified, namely - the language should not be standardized and it must have been created after 1970.

\n\n

If we want our resultset to satisfy any one of our conditions, we use the OR operator.

\n\n

Listing: using the OR operator

\n\n

1 SELECT language, author, year FROM proglang_tbl\n2 WHERE year > 1970 OR standard IS NULL;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	author	year
Prolog	Colmerauer	1972
Perl	Wall	1987
Tcl	Ousterhout	1988

\n\n

The result now contains languages which were created either after 1970 like Prolog or which are not standardized like Perl or Tcl. In this particular example, the first condition satisfies all the rows of the resultset. But if there were a language which was created before 1970 and wasn’t yet standardized, it would show up as a result of this query. We can even create yet more complex queries by combining these operators.

\n\n\n

\n6. Manipulating Data

\n\n

In this chapter we study the Data Manipulation Language (DML) part of SQL which is used to make changes to the data inside a relational database. The three basic commands of DML are as follows.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

INSERT	Populates tables with new data
UPDATE	Updates existing data
DELETE	Deletes data from tables

\n\n

We have already seen a few examples on the INSERT statement including simple inserts and selective field insertions. Thus we will concentrate on other ways to use this statement.

\n\n

\n6.1 Inserting NULL’s

\n\n

In previous chapters, we have seen that not specifying a column value while doing selective field insert operations results in a null value being set for them. We can also explicitly use the keyword NULL in SQL to signify null values in statements like INSERT.

\n\n

Listing: Inserting NULL values

\n\n

1 INSERT INTO proglang_tbl VALUES (4, 'Tcl', 'Ousterhout', '1988', NULL);\n

\n\n\n\n

Running a query to show the contents of the entire table helps us to verify the result.

\n\n

1 SELECT * FROM proglang_tbl;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a table with NULL values
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI
4	Tcl	Ousterhout	1988	(null)

\n\n

\n6.2 Inserting data into a table from another table

\n\n

You can insert new records into a table from another one by using a combination of INSERT and SELECT. Since a query would return you some records, combining it with an insertion command would enter these records into the new table. You can even use a WHERE conditional to limit or filter the records you wish to enter into the new table. We will now create a new table called stdlang_tbl, which will have only two fields - language and standard. In this we would insert rows from the proglang_tbl table which have a non-null value in the standard field. This will also demonstrate our first use of a boolean operator - NOT.

\n\n

Listing: Using INSERT and SELECT to conditionally load data into another table

\n\n

1 CREATE TABLE stdlang_tbl (language varchar(20), standard varchar (10));\n2 \n3 INSERT INTO stdlang_tbl SELECT language, standard FROM proglang_tbl WHERE standa\\\n4 rd IS NOT NULL;\n

\n\n\n\n

When you view the contents of this table, you will notice that it has picked up the two languages which actually had a standard\ncolumn value.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: Contents of the stdlang_tbl table
language	standard
Prolog	ISO
APL	ANSI

\n\n

\n6.3 Updating existing data

\n\n

To modify some data in a record, we use the UPDATE command. While it cannot add or delete records (those responsibilities are delegated to other commands), if a record exists it can modify its data even affecting multiple fields in one go and applying conditions. The general syntax of an UPDATE statement is given below.

\n\n

Listing: General Syntax of the UPDATE command

\n\n

1 UPDATE <table_name> SET\n2  <column1> = <value>,\n3  <column2> = <value>,\n4  <column3> = <value>\n5  . . .\n6 WHERE <condition>;\n

\n\n\n\n

Let us now return to our proglang_tbl table and add a new row about the Forth programming language.

\n\n

1 INSERT INTO proglang_tbl VALUES (5, 'Forth', 'Moore', 1973, NULL);\n

\n\n\n\n

We later realize that the language actually was created near 1972 (instead of 1973) and it actually has been standardized in 1994 by the ANSI. Thus we write our update query to reflect the same.

\n\n

Listing: Updating multiple fields in a single statement

\n\n

1 UPDATE proglang_tbl SET year = 1972,\n2   standard = 'ANSI' WHERE language = 'Forth';\n

\n\n\n\n

If you’ve typed the statement correctly and no errors are thrown back, the contents of the record in question would have been modified as intended. Verifying the result of the same involves a simple query the likes of which we have seen in previous examples.

\n\n

\n6.4 Deleting data from tables

\n\n

You can use the DELETE command to delete records from a table. This means that you can choose which records you want to delete based on a condition, or delete all records but you cannot delete certain fields of a record using this statement. The general syntax of the DELETE statement is given below.

\n\n

Listing: General syntax of DELETE

\n\n

1 DELETE FROM <table_name> WHERE <condition>;\n

\n\n\n\n

While putting a conditional clause in the DELETE is optional, it is almost always used. Simply because not using it would cause all the records to be deleted from a table, which is a rarely valid need. We now write the full statement to delete the record corresponding to Forth from the table.

\n\n

Listing: Deleting a record from the proglang_tbl table

\n\n

1 DELETE FROM proglang_tbl WHERE language = 'Forth';\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: table contents after the record deletion
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI
4	Tcl	Ousterhout	1988	(null)

\n\n\n

\n7. Organizing your data

\n\n

The number of fields you wish to store in your database would be a larger value than the five column table we saw earlier chapters. Also, some assumptions were made intrinsically on the kind of data we will store in the table. But this is not always the case in real life. In reality the data we encounter will be complex, even redundant. This is where the study of data modelling techniques and database design come in. While it is advised that the reader refer to a more comprehensive treatise on this subject, nonetheless we will try to study some good relational database design principles since the study would come in handy while learning SQL statements for multiple tables.

\n\n

\n7.1 Normalization

\n\n

Let us suppose we have a database of employees in a fictional institution as given below. If the database structure has not been modelled but has been extracted from a raw collection of information available, redundancy is expected.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: the fictional firm’s database
employee_id	name	skill	manager_id	location
1	Socrates	Philosophy	(null)	Greece
2	Plato	Writing	1	Greece
3	Aristotle	Science	2	Greece
4	Descartes	Philosophy	(null)	France
4	Descartes	Philosophy	(null)	Netherlands

\n\n

We can see that Descartes has two rows because he spent his life in both France and Netherlands. At a later point we decide that we wish to classify him with a different skill, we would have to update both rows since they should contain an identical (primary) skill. It would be easier to have a separate table for skills and and somehow allow the records which share the same skill to refer to this table. This way if we wish to reflect that both Socrates and Descartes were thinkers in Western Philosophy renaming the skill record in the second table would do the trick.

\n\n

This process of breaking down a raw database into logical tables and removing redundancies is called Normalization. There are even levels of normalization called normal forms which dictate on how to acheive the desired design.

\n\n

\n7.2 Atomicity

\n\n

In the programming language examples we’ve seen, our assumption has always been that a language has a single author. But there are countless languages where multiple people contributed to the core design and should rightfully be acknowledged in our table. How would we go about making such a record? Let us take the case of BASIC which was designed by John Kemeny and Thomas Kurtz. The easiest option to add this new record into the table is to fit both author’s in the author field.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a record with a non-atomic field value
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI
4	Tcl	Ousterhout	1988	(null)
5	BASIC	Kemeny, Kurtz	1964	ANSI

\n\n

You can immediately see that it would be difficult to write a query to retrieve this record based on the author field. If the data written as “Kemeny, Kurtz” or “Kurtz, Kemeny” or even “Kemeny & Kurtz”, it would be extremely difficult to put the right string in the WHERE conditional clause of the query. This is often the case with multiple values, and the solution is to redesign the table structure to make all field value atomic.

\n\n

\n7.3 Repeating Groups

\n\n

Another simple (but ultimately wrong) approach that comes to mind is to split the author field into two parts - author1 and author2. If a language has only one author, the author2 field would contain a null value. Can you spot the problem that will arise from this design decision?

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a table with a repeating group
id	language	author1	author2	year	standard
1	Prolog	Colmerauer	(null)	1972	ISO
2	Perl	Wall	(null)	1987	(null)
3	APL	Iverson	(null)	1964	ANSI
4	Tcl	Ousterhout	(null)	1988	(null)
5	BASIC	Kemeny	Kurtz	1964	ANSI

\n\n

This imposes an artificial constraint on how many authors a language can have. It seems to work fine for a couple of them, but what if a programming language was designed by a committee of a dozen or more people? At the database design time, how do we fix the number of authors we wish to support? This kind of design is referred to as a repeating group and must be actively avoided.

\n\n

\n7.4 Splitting the table

\n\n

The correct design to remove the problems listed above is to split the table into two - one holding the author details and one detailing the language.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a table holding author details
author_id	author	language_id
1	Colmerauer	1
2	Wall	2
3	Ousterhout	4
4	Iverson	3
5	Kemeny	5
6	Kurtz	5

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: a table holding programming language details
id	language	year	standard
1	Prolog	1972	ISO
2	Perl	1987	(null)
3	APL	1964	ANSI
4	Tcl	1988	(null)
5	BASIC	1964	ANSI

\n\n

Once you have removed the non-atomicity of fields and repeating groups alongwith assigning unique id’s to your tables, your table structure is now in the first normal form. The author table’s language_id field which refers to the id field of the language table is called a foreign key constraint.

\n\n

Listing: creating the new programming languages table

\n\n

1 CREATE TABLE newlang_tbl (id       INTEGER     NOT NULL PRIMARY KEY,\n2                           language VARCHAR(20) NOT NULL,\n3                           year     INTEGER     NOT NULL,\n4                           standard VARCHAR(10) NULL);\n

\n\n\n\n

Listing: creating the authors table

\n\n

1 CREATE TABLE authors_tbl (author_id   INTEGER     NOT NULL,\n2                           author      VARCHAR(25) NOT NULL,\n3                           language_id INTEGER REFERENCES newlang_tbl(id));\n

\n\n\n\n

Notice that in the author’s table we’ve made a foreign key constraint by making the language_id field reference the id field of the new programming languages table using the keyword REFERENCES. You can only create a foreign key reference a primary or unique key, otherwise during the constraint creation time we would recieve an error similar to the following.

\n\n

1 E_PS0490 CREATE/ALTER TABLE: The referenced columns in table 'newlang_tbl'\n2     do not form a unique constraint; a foreign key may only reference\n3     columns in a unique or primary key constraint.\n4     (Thu May 17 15:28:45 2012)\n

\n\n\n\n

Since we have created a reference to the language_id, inserting a row in the author’s table which does not yet have a language entry would also result in an error, called a Referential Integrity violation.

\n\n

1 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (5, 'Kemeny', 5)\n2 \n3 E_US1906 Cannot INSERT into table '\"authors_tbl\"' because the values do\n4     not match those in table '\"newlang_tbl\"' (violation of REFERENTIAL\n5     constraint '\"$autho_r0000010c00000000\"').\n

\n\n\n\n

However when done sequentially, i.e. the language first and then its corresponding entry in the author table, everything works out.

\n\n

Listing: making entries for BASIC in both the tables

\n\n

1 INSERT INTO newlang_tbl (id, language, year, standard) VALUES (5, 'BASIC', 1964,\\\n2  'ANSI');\n3 \n4 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (5, 'Kemeny', 5);\n

\n\n\n\n

The other statements to get fully populated tables are given below.

\n\n

 1 INSERT INTO newlang_tbl  (id, language, year,  standard) VALUES (1, 'Prolog', 19\\\n 2 72, 'ISO');\n 3 INSERT INTO newlang_tbl  (id, language, year) VALUES (2, 'Perl', 1987);\n 4 INSERT INTO newlang_tbl  (id, language, year,  standard) VALUES (3, 'APL', 1964,\\\n 5  'ANSI');\n 6 INSERT INTO newlang_tbl  (id, language, year) VALUES (4, 'Tcl', 1988);\n 7 \n 8 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (6, 'Kurtz', 5);\n 9 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (1, 'Colmerauer'\\\n10 , 1);\n11 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (2, 'Wall', 2);\n12 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (3, 'Ousterhout'\\\n13 , 4);\n14 INSERT INTO authors_tbl (author_id, author, language_id) VALUES (4, 'Iverson', 3\\\n15 );\n

\n\n\n\n\n

\n8. Doing more with queries

\n\n

We have already seen some basic queries, how to order the results of a query and how to put conditions on the query output. Let us now see more examples of how we can modify our SELECT statements to suit our needs.

\n\n

\n8.1 Counting the records in a table

\n\n

Sometimes we just wish to know how many records exist in a table without actually outputting the entire contents of these records. This can be achieved through the use of a SQL function called COUNT. Let us first see the contents of the proglang_tbl table.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: contents of our programming languages table
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)
3	APL	Iverson	1964	ANSI
4	Tcl	Ousterhout	1988	(null)

\n\n

Listing: Query to count number of records in the table

\n\n

1 SELECT COUNT(*) FROM proglang_tbl;\n

\n\n\n\n

The output returned will be a single record with a single field with the value as 4. The function COUNT took one argument i.e. what to count and we provided it with * which means the entire record. Thus we achieved our purpose of counting records in a table.

\n\n

What would happen if instead of giving an entire record to count, we explicitly specify a column? And what if the column had null values? Let’s see this scenario by counting on the standard field of the table.

\n\n

Listing: Query to count number of standard field values in the table

\n\n

1 SELECT COUNT(standard) FROM proglang_tbl;\n

\n\n\n\n

The output in this case would be the value 2, because we only have two records with non-null values in the standard field.

\n\n

\n8.2 Column Aliases

\n\n

Queries are frequently consumed directly as reports since SQL provides enough functionality to give meaning to data stored inside a RDBMS. One of the features allowing this is Column Aliases, which let you rename column headings in the resultant output. The general syntax for creating a column alias is given below.

\n\n

Listing: General Syntax for creating column aliases

\n\n

1 SELECT <column1> <alias1>, <column2> <alias2> ... from <table>;\n

\n\n\n\n

For example, we wish to output our programming languages table with a few columns only. But we do not wish to call the authors of the language as authors. The person wanting the report wishes they be called creators. This can be simply done by using the query below.

\n\n

Listing: Renaming the author field to creator for reporting purposes

\n\n

1 SELECT id, language, author creator from proglang_tbl;\n

\n\n\n\n

While creating a column alias will not permanantly rename a field, it will show up in the resultant output.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: the column alias output
id	language	creator
1	Prolog	Colmerauer
2	Perl	Wall
3	APL	Iverson
4	Tcl	Ousterhout

\n\n

\n8.3 Order of execution of SELECT queries

\n\n

A query is not evaluated from left to right, there is a specific sequence in which its various parts are evaluated as given below.

\n\n

FROM clause
WHERE clause
GROUP BY clause
HAVING clause
SELECT clause
ORDER BY clause

\n\n

There is an interesting corollary of having the SELECT evaluation being lower (read later) than the WHERE clause. Most database\nmanagement systems, like Microsoft SQL Server will not allow you to use a column alias in the filtering conditions. So a query like\nthe one given below would not work.

\n\n

Listing: using a column alias in the WHERE clause

\n\n

1 SELECT author Scientist FROM authors_tbl WHERE Scientist = 'Wall';\n

\n\n\n\n

However, this works fine in SQLite. Yet another example of how there exists subtle differences between DBMSs’.

\n\n

\n8.4 Using the LIKE operator

\n\n

While putting conditions on a query using WHERE clauses, we have already seen comparison operators = and IS NULL. Now we take a look at the LIKE operator which will help us with wildcard comparisons. For matching we are provided with two wilcard characters to use with LIKE.

\n\n\n \n \n \n \n \n \n \n \n \n \n\n

1) % (Percent)	Used to match multiple characters including a single character and no character
2) _ (Underscore)	Used to match exactly one character

\n\n

We will first use the % character for wildcard matching. Let us suppose we wish to list out languages that start with the letter P.

\n\n

Listing: using the LIKE operator and % wildcard

\n\n

1 SELECT * FROM proglang_tbl WHERE language LIKE 'P%';\n

\n\n\n\n

The output of the above query should be all language records whose name begins with the letter capital P. Note that this would not include any language that starts with the small letter p.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: all languages starting with P
id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	(null)

\n\n

We can see that using the % wildcard allowed us to match multiple characters like erl in the case of Perl. But what if we wanted to restrict how many characters we wished to match? What if our goal was to write a query which displays the languages ending in the letter l, but are only 3 characters in length? The first condition could have been satisfied using the pattern %l, but to satisfy both conditions in the same query we use the _ wildcard. A pattern like %l would result in returning both Perl and Tcl but we modify our pattern suitably to return only the latter.

\n\n

1 SELECT * FROM proglang_tbl WHERE language LIKE '__l';\n

\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: output for _ wildcard matching
id	language	author	year	standard
4	Tcl	Ousterhout	1988	(null)

\n\n

Note that the result did not include Perl since we explicitly gave two underscores to match 2 characters only. Also it did not match APL since SQL data is case sensitive and l is not equal to L.

\n\n\n

\n9. Calculated Fields

\n\n

We have already seen column aliases which allow us to rename a field’s name in the query output. But we frequently encounter conditions which require changes to a field value. This is where the concept of a calculated field comes in.

\n\n

\n9.1 Mathematical calculations

\n\n

Any numeric field can be operated upon by mathematical operators we are all familiar with. We can add, subtract, multiply, divide and even find the remainder of a division operation fairly easily. While the operators supported differ in various implementations, the ones given below should be available across any RDBMS you come across.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Addition	+
Subtraction	-
Multiplication	*
Division	/
Remainder	%

\n\n

Let us take our programming languages table and try to find out the decade in which the language was created. For example, Prolog was created in the 1970’s decade. Let us try to find out this fact from the year of creation available to us. One approach is to find the remainder of the year when divided by 10, which is the number of years in a decade. This is the value that specifies how many years has it been since the start of that decade.

\n\n

1 select language, (year%10) remain from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	remain
Prolog	2
Perl	7
APL	4
Tcl	8

\n\n

Now if we subtract this value from the year of creation itself, we would get the decade in which the programming language was created.

\n\n

1 select language, year-(year%10) decade from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	decade
Prolog	1970
Perl	1980
APL	1960
Tcl	1980

\n\n

Another approach is to divide the year by 10 and then multiply it by 10. This is slightly less straightforward because it relies on the definition of the integer data type. Since an integer cannot store decimal points, division by ten would silently chop off the remainder. 1972 divided by 10 would be 197 discarding the .2 bit. If we multiply this value by 10, we would get our desired decade value.

\n\n

1 select language, (year/10)*10 decade from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	decade
Prolog	1970
Perl	1980
APL	1960
Tcl	1980

\n\n

\n9.2 String operations

\n\n

By far the most commonly used string operation is concatenation. It means to join or combine strings. However since even numeric fields can be treated as a string, we can use the concatenation operator || on them too. See the example below to modify our decade field to include some characters.

\n\n

1 select language, 'The '||((year/10)*10)||'s' decade from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	decade
Prolog	The 1970s
Perl	The 1980s
APL	The 1960s
Tcl	The 1980s

\n\n

Note that the concatenation operator manifests itself in different forms in different implementations. SQLite and Oracle use the shown || symbols whereas Ingres, MySQL and Microsoft SQL Server use + to denote concatenation. Their effect however is the same.

\n\n

\n9.3 Literal Values

\n\n

There are cases when one needs to use a fixed literal value as the values of a new column. Like column aliases can change the column header for readability, literal values change record values. In a sense they are not calculated fields, but fixed fields inserted in specific positions of a record. An example will help illustrate this - supposing to wish to really clarify that the year of language creation, as not just a number but also to include the characters AD

\n\n

1 select language, year, 'AD' from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	year	‘AD’
Prolog	1972	AD
Perl	1987	AD
APL	1964	AD
Tcl	1988	AD

\n\n

We can even use numeric literal values the same way, omitting the quotation marks for such values. A common utility for literal values arises when the user has to copy-paste data from their database query output into another tool like a spreadsheet or wordprocessor.

\n\n\n

\n10. Aggregation and Grouping

\n\n

\n10.1 Aggregate Functions

\n\n

An aggregate function is used to compute summarization information from a table or tables. We have already seen the COUNT aggregate function which counts the records matched. Similarly there are other aggregation functions in SQL like AVG for calculating averages, SUM for computing totals and MAX, MIN for finding out maxima and minima values respectively.

\n\n

\n10.2 Using DISTINCT with COUNT

\n\n

We have already seen the COUNT function, but we can further control its output using the optional argument DISTINCT. This allows us to count only non-duplicate values of the input specified. To illustrate this concept, we will now insert some rows into our proglang_tbl table.

\n\n

Listing: Inserting some new rows in our programming languages table

\n\n

1 INSERT INTO proglang_tbl (id, language, author, year, standard) VALUES (5, 'Fort\\\n2 ran', 'Backus', 1957, 'ANSI');\n3 \n4 INSERT INTO proglang_tbl (id, language, author, year, standard) VALUES (6, 'PL/I\\\n5 ', 'IBM', 1964, 'ECMA');\n

\n\n\n\n

Note the new data choice that we are populating. With Fortran we are adding a new programming language that has a standard by the ANSI. With PL/I we now have a third distinctive standards organisation - ECMA. PL/I also shares the same birth year as APL (1964) giving us a duplicate year field. Now let us run a query to check how many distinct year and standard values we have.

\n\n

Listing: Counting distinct year values

\n\n

1 SELECT COUNT (DISTINCT year) FROM proglang_tbl;\n2 \n3 > 5\n

\n\n\n\n

Listing: Counting distinct standard values

\n\n

1 SELECT COUNT (DISTINCT standard) FROM proglang_tbl;\n2 \n3 > 3\n

\n\n\n\n

The first query result is straightforward. We have 6 rows but two of them share a common year value, thus giving us the result 5. In the second result, out of 6 rows only 4 of them have values. Two rows have a NULL value in them meaning those languages have no standard. Among the 4 rows, two of them share a common value, giving us the result - 3. Note that the DISTINCT clause did not count NULL values as truly distinct values.

\n\n

\n10.3 Using MIN to find minimum values

\n\n

The MIN function is fairly straightforward. It looks at a particular set of rows and finds the minimum value of the column which is provided as an argument to it. For example, in our example table we wish to find out from which year do we have records of programming languages. Analyzing the problem at hand, we see that if we apply the aggregate function MIN to the field year in our table, we should get the desired output.

\n\n

Listing: finding out the earliest year value in our table

\n\n

1 SELECT MIN(year) from proglang_tbl;\n2 \n3 > 1957\n

\n\n\n\n

The MAX function similarly finds the largest value in the column provided to it as an argument.

\n\n

Listing: finding out the latest year value in our table and the programming language associated with it

\n\n

1 select language, MAX(year) year from proglang_tbl;\n

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n\n

language	year
1988	Tcl

\n\n

\n10.4 Grouping Data

\n\n

The GROUP BY clause of a SELECT query is used to group records based upon their field values. This clause is placed after the WHERE conditional. For example, in our sample table we can group data by which committee decided on their standard.

\n\n

Listing: Grouping records by its fields

\n\n

1 SELECT language, standard FROM proglang_tbl\n2 WHERE standard IS NOT NULL\n3 GROUP BY standard, language;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: output for grouping records
language	standard
APL	ANSI
Fortran	ANSI
PL/I	ECMA
Prolog	ISO

\n\n

The interesting thing to note here is the rule that the columns listed in the SELECT clause must be present in the GROUP BY clause. This leads us to the following two corollaries.

\n\n

You cannot group by a column which is not present in the SELECT list.
You must specify all the columns in the grouping clause which are present in the SELECT list.

\n\n

Another useful way to use grouping is to combine the operation with an aggregate function. Supposing we wish to count how many standards a particular organization has in our table. This can be achieved by combining the GROUP BY clause with the COUNT aggregate function as given below.

\n\n

Listing: using GROUP BY with aggregate functions

\n\n

1 SELECT standard, count(*) FROM proglang_tbl GROUP BY standard; \n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: query output showing the count of standard organizations in our table
standard	col2
ANSI	2
ECMA	1
ISO	1
(null)	2

\n\n

\n10.5 The HAVING Clause

\n\n

Like a WHERE clause places conditions on the fields of a query, the HAVING clause places conditions on the groups created by GROUP BY. It must be placed immediately after the GROUP BY but before the ORDER BY clause.

\n\n

Listing: demonstration of the HAVING clause

\n\n

1 SELECT language, standard, year FROM proglang_tbl\n2 GROUP BY standard, year, language HAVING year < 1980;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: output of the HAVING clause demonstration query
language	standard	year
APL	ANSI	1964
Fortran	ANSI	1957
PL/I	ECMA	1964
Prolog	ISO	1972

\n\n

From the output we can clearly see that the records for Perl and Tcl are left out since they do not satisfy the HAVING conditional of being created before 1980.

\n\n\n\n\n

\n11. Understanding Joins

\n\n

\n11.1 What is a Join?

\n\n

A join operation allows you to retrieve data from multiple tables in a single SELECT query. Two tables can be joined by a single join operator, but the result can be joined again with other tables. There must exist a same or similar column between the tables being joined.

\n\n

When you design an entire database system using good design principles like normalization, we often require the use of joins to give a complete picture to a user’s query. For example, we split our programming languages table into two - one holding the author details and the other holding information about the languages itself. To show a report listing authors and which programming language they created, we would have to use a join.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: authors_tbl contents
author_id	author	language_id
1	Colmerauer	1
2	Wall	2
3	Ousterhout	4
4	Iverson	3
5	Kemeny	5
6	Kurtz	5

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: newlang_tbl contents
id	language	year	standard
1	Prolog	1972	ISO
2	Perl	1987	(null)
3	APL	1964	ANSI
4	Tcl	1988	(null)
5	BASIC	1964	ANSI

\n\n

We now form a query to show our desired output - the list of all authors with the corresponding language they developed. We choose our join column as the language_id field from the authors table. This corresponds to the id field in the languages table.

\n\n

Listing: running a join operation on our two tables

\n\n

1 SELECT author, language FROM authors_tbl, newlang_tbl\n2 WHERE language_id = id;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: result of our join query
author	language
Colmerauer	Prolog
Wall	Perl
Iverson	APL
Ousterhout	Tcl
Kemeny	BASIC
Kurtz	BASIC

\n\n

The output of our query combines a column from both tables giving us a better report. The language_id = id is called the join condition. Since the operator used in the join condition is an equality operator (=), this join is called as an equijoin. Another important thing to note is that the columns participating in the join condition are not the ones we choose to be in the result of the query.

\n\n

\n11.2 Alternative Join Syntax

\n\n

You would have noticed that we formed our join query without much special syntax, using our regular FROM/WHERE combination. The SQL-92 standard introduced the JOIN keyword to allow us to form join queries. Since it was introduced earlier, the FROM/WHERE syntax is more common. But now that the majority of database vendors have implemented most of the SQL-92 standard, the JOIN syntax is also in widespread use. Below is the JOIN syntax equivalent of the query we just wrote to display which author created which programming language.

\n\n

Listing: Rewriting our query using the JOIN(SQL-92) syntax

\n\n

1 SELECT author, language FROM authors_tbl JOIN newlang_tbl\n2 ON language_id = id;\n

\n\n\n\n

Notice that instead separating the two tables using a comma (thereby making it a list), we use the JOIN keyword. The columns which participate in the join condition are preceded by the ON keyword. The WHERE clause can then be used after the join condition specification (ON clause) to specify any further conditions if needed.

\n\n

\n11.3 Resolving ambiguity in join columns

\n\n

In our example the join condition fields had distinct names - id and language_id. But what if in our languages table (newlang_tbl) we kept the key field’s name as language_id. This would create an ambiguity in the join condition, which would become the confusing language_id = language_id. To resolve this, we need to qualify the column by prepending it by the table name it belongs to and a .(period).

\n\n

Listing: Resolving the naming ambiguity by qualifying the columns

\n\n

1 SELECT author, language FROM authors_tbl JOIN newlang_tbl\n2 ON authors_tbl.language_id = newlang_tbl.language_id;\n

\n\n\n\n

Another way to solve such ambiguity is to qualify the columns using table aliases. The concept is to give a short name to a table and then use this to qualify the columns instead of a long, unwieldy table name.

\n\n

Listing: using table aliases

\n\n

1 SELECT author, language FROM authors_tbl a JOIN newlang_tbl l\n2 ON a.language_id = l.id;\n

\n\n\n\n

Here the authors table is given the alias a and the languages table is given the alias l. It is generally considered a good practice to qualify column names of a join condition regardless of whether there is a name ambiguity or not.

\n\n

\n11.4 Cross Joins

\n\n

You might think what would happen if we left out the join condition from our query. Well what happens in the background of running a join query is that first all possible combinations of rows are made from the tables participating in the join. Then the rows which satisfy the join condition are chosen for the output (or further processing). If we leave out the join condition, we get as the output all possible combinations of records. This is called a Cross Join** or Cartesian Product of the tables usually denoted by the sign X.

\n\n

Listing: query for showing the cartesian product of our tables

\n\n

1 SELECT author, language FROM authors_tbl, newlang_tbl;\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: the cartesian product of our tables
author	language
Kemeny	BASIC
Kurtz	BASIC
Colmerauer	BASIC
Wall	BASIC
Ousterhout	BASIC
Iverson	BASIC
Kemeny	Prolog
Kurtz	Prolog
Colmerauer	Prolog
Wall	Prolog
Ousterhout	Prolog
Iverson	Prolog
Kemeny	Perl
Kurtz	Perl
Colmerauer	Perl
Wall	Perl
Ousterhout	Perl
…	…

\n\n

Another way to rewrite this query is to actually use the JOIN keyword with a preceding argument CROSS as shown below.

\n\n

Listing: rewriting the query using CROSS JOIN

\n\n

1 SELECT author, language FROM authors_tbl CROSS JOIN newlang_tbl;\n

\n\n\n\n

\n11.5 Self Joins

\n\n

Sometimes a table within its own columns has meaningful data but one (or more) of its fields refer to another field in the same table. For example if we have a table in which we capture programming languages which influenced other programming languages and denote the influence relationship by the language id, to show the resolved output we would have to join the table with itself. This is also called a SELF JOIN. Consider the table created below and pay close attention to the data being inserted.

\n\n

Listing: creating and populating our language influence table

\n\n

 1 CREATE TABLE inflang_tbl (id            INTEGER     PRIMARY KEY,\n 2                           language      VARCHAR(20) NOT NULL,\n 3                           influenced_by INTEGER);\n 4 \n 5 INSERT INTO inflang_tbl (id, language)\n 6  VALUES (1, 'Fortran');\n 7 \n 8 INSERT INTO inflang_tbl (id, language, influenced_by)\n 9 VALUES (2, 'Pascal', 3);\n10 \n11 INSERT INTO inflang_tbl (id, language, influenced_by)\n12 VALUES (3, 'Algol', 1);\n

\n\n\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: contents of inflang_tbl
id	language	influenced_by
1	Fortran	(null)
2	Pascal	3
3	Algol	1

\n\n

Our goal is to now write a self join query to display which language influenced which one, i.e. resolve the influenced_by column.

\n\n

Listing: running a self join query

\n\n

1 SELECT l1.language, l2.language AS influenced\n2 FROM inflang_tbl l1, inflang_tbl l2\n3 WHERE l1.id = l2.influenced_by;\n

\n\n\n\n

Notice the use of table aliases to qualify the join condition columns as separate and the use of the AS keyword which renames the column in the output.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: result of our self join query
language	influenced
Algol	Pascal
Fortran	Algol

\n\n\n

\n12. Subqueries

\n\n

A subquery, simply put, is a query written as a part of a bigger statement. Think of it as a SELECT statement inside another \none. The result of the inner SELECT can then be used in the outer query. Let us take a simple example to illustrate this. \nConsider the same source tables as the ones in the Understanding Joins chapter - authors_tbl and newlang_tbl. We will try \nto write a query (and a subquery) to display the author of a particular language.

\n\n

Listing: A simple subquery example

\n\n

1 SELECT author FROM authors_tbl \n2   WHERE language_id IN (SELECT id FROM newlang_tbl WHERE language='Tcl');\n3 \n4 > Ousterhout\n

\n\n\n\n

The subquery SELECT id FROM newlang_tbl WHERE language='Tcl' picks the correct language id from the newlang_tbl and passes it on\nto the outer query on the authors table. This frees us from the responsibility of joining the two tables using the language id\nfield. Which approach to take in certain situations - a join, a subquery or a combination of both - is mostly a matter of personal \npreference. Other times, one approach will be clearly the superior choice.

\n\n

\n12.1 Types of subqueries

\n\n

We can broadly classify subqueries into three categories.

\n\n

\nScalar subqueries\nA subquery that returns only a single column of a single row as its output. The example in the previous section, where the subquery\nreturns the id for Tcl is a scalar subquery.
\nTable subqueries\nA table subquery can return more than a single row and many columns per row. In essence, it can return a table itself to take part\nin your outer query. Let us take an example where we wish to display all the programming language writers who created a language\nafter 1980.

\n\n

Listing: A table subquery example

\n\n

1 SELECT author, language FROM authors_tbl a, (SELECT id, language FROM newlang_tb\\\n2 l WHERE year > 1980) n \n3   WHERE a.language_id = n.id;\n

\n\n\n\n

Carefully study the FROM clause of the query above. Our table subquery is placed within it and it returns a set of languages which\nwere created after 1980. The result consists of two rows and two columns, one of which i.e. language is picked up to be displayed\nin the final output.

\n\n\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Figure: authors who created their languages after 1980
author	language
Wall	Perl
Ousterhout	Tcl

\n\n

\nRow subqueries\nA subquery that returns a single row but more than one column is called a row subquery. These are the least important type of\nsubqueries since most database management systems do not support it, including SQLite.

\n\n

\n12.2 Using subqueries in INSERT statements

\n\n

We can even use subqueries inside other SQL statement like INSERT. Let us try to add a new language and a new author in our tables\nand ease our task of remembering id numbers by just a bit by using subqueries.

\n\n

Listing: Inserting a new programming language

\n\n

1 INSERT INTO newlang_tbl (id, language, year, standard) \n2   VALUES (6, 'Pascal', 1970, 'ISO');\n

\n\n\n\n

The contents of our table now look as shown below.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

id	language	year	standard
1	Prolog	1972	ISO
2	Perl	1987
3	APL	1964	ANSI
4	Tcl	1988
5	BASIC	1964	ANSI
6	Pascal	1970	ISO

\n\n

While inserting a new entry into the authors_tbl, we can either remember that we used the language_id as 6 for Pascal, or use a\nsubquery. Let us see an example of the latter approach.

\n\n

Listing: Inserting a new author using a subquery

\n\n

1 INSERT INTO authors_tbl (author_id, author, language_id)\n2   VALUES (7, 'Wirth', (SELECT id FROM newlang_tbl WHERE language='Pascal'));\n

\n\n\n\n

We believe that this should put the correct language id for Wirth since he created Pascal. Let us verify this belief by looking at\nte contents of the table.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

author_id	author	language_id
5	Kemeny	5
6	Kurtz	5
1	Colmerauer	1
2	Wall	2
3	Ousterhout	4
4	Iverson	3
7	Wirth	6

\n\n\n

\n13. Working with Sets

\n\n

Set theory is a branch of discrete mathematics that deals with a collection of objects. There is a lot of conceptual overlap\nbetween set theory and relational database concepts. It is no wonder that the output of a query is frequently called a resultset.\nPrimitive set theoretic operations like union, intersection and difference are increasingly supported in various SQL\nimplementations. We will now explore the theory behind these operations and how to use them in SQL.

\n\n

\n13.1 Union

\n\n

The union is an operation that combines elements of two sets. Let’s say we have the following two sets:

\n\n

1 set1 = { 1, 3, 5}\n2 \n3 set2 = { 1, 2, 3}\n

\n\n\n\n

The resulting union set will be a set consisting of all of these elements repeated exactly once, i.e. no duplicates are allowed. \nNote that the order of a set is unimportant. Think of it as a bag of elements rather than an ordered collection.

\n\n

1 set1 UNION set2 = { 1, 3, 5, 2 }\n

\n\n\n\n

Let’s now look at how to use simulate the UNION operation in SQL. Consider the table and its data below.

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

id	language	author	year	standard
1	Prolog	Colmerauer	1972	ISO
2	Perl	Wall	1987	null
3	APL	Iverson	1964	ANSI
4	RPG	IBM	1964	ISO

\n\n

If we wanted to get the list of creation years of languages standardized by either ANSI or ISO, we could use a UNION keyword to\nachieve this.

\n\n

Listing: using a UNION operator

\n\n

1 SELECT year FROM proglang_tbl WHERE standard='ANSI'\n2 UNION\n3 SELECT year FROM proglang_tbl WHERE standard='ISO'\n4 \n5 > year\n6   ----\n7   1964\n8   1972\n

\n\n\n\n

The result is expectedly two rows because 1964 is a repeated year being the creation year of two languages which were standardized.\nWe can get three distinct rows with the year 1964 being repeated twice by using the UNION ALL operator instead of simply UNION.\nThe rest of the query remains exactly the same.

\n\n

\n13.2 Intersection

\n\n

The intersection operation that outputs only the common elements in the input sets. If we apply an intersection to the two sets\nin the previous section we get a resulting set of two elements.

\n\n

1 set1 INTERSECTION set2 = { 1, 3 }\n

\n\n\n\n

Translating this to SQL is pretty simple; instead of using UNION we use the keyword INTERSECT to get common elements.

\n\n

Listing: using the INTERSECT

\n\n

1 SELECT year FROM proglang_tbl WHERE standard='ANSI'\n2 INTERSECT\n3 SELECT year FROM proglang_tbl WHERE standard='ISO'\n4 \n5 > year\n6   ----\n7   1964\n

\n\n\n\n

Note that while the ANSI SQL standard does provision for an INTERSECT ALL operator, I’m yet to come across a database management\nsystem that implements it.

\n\n

\n13.3 Difference

\n\n

The difference operation between sets, written as set1 - set2 is a list of all elements in set1 that do not occur in set2.\nIf an element is only in set2, it will not be captured by the plain difference operation.

\n\n

1 set1 DIFFERENCE set2 = { 5 }\n2 set2 DIFFERENCE set1 = { 2 }\n

\n\n\n\n

Let’s try and write a SQL statement to emulate this logic with our familiar IN and NOT IN operators. Suppose we wish to list\nout the years of creation of languages which were standardized by ISO but not the ANSI. From our source table we find that two\nlanguages were standardized with years 1972 and 1964. But since in 1964, APL was created which was eventually standardized by\nANSI, we should ideally be left with the answer 1972.

\n\n

Listing: trying to write set difference with IN

\n\n

1 SELECT year FROM proglang_tbl \n2   WHERE standard IN ('ISO') \n3     AND standard NOT IN ('ANSI');\n4 \n5 > year\n6   ----\n7   1972\n8   1964\n

\n\n\n\n

Whoa, what sorcery is this!?! We thought 1964 would be ineligible because of ANSI standardization. But clearly this is not the\ncase. What has happened actually is that first there was a scan of ISO rows - giving us two values. Then ANSI rows were discounted\nbut not necessarily from the first result but the table as a whole. So while the APL 1964 was left off, the RPG 1964 still\nremained, effectively making our second condition worthless. The correct way to achieve this is using the set difference operator\nEXCEPT as below.

\n\n

Listing: set difference with EXCEPT

\n\n

1 SELECT year FROM proglang_tbl WHERE standard IN ('ISO')\n2 EXCEPT\n3 SELECT year FROM proglang_tbl WHERE standard IN ('ANSI')\n4 \n5 > year\n6   ----\n7   1972\n

\n\n\n\n

Voila, this seems to yield the correct answer! When we write more than a single SELECT as a part of a single query and join them\nusing an set theoretic operator, such statements are called compound queries.

\n\n\n\n

Appendix: Major Database Management Systems

\n\n

Ingres (Actian Corporation)\n
A full featured relational database management system available as a proprietary or an open source edition.
\n\n
http://www.actian.com/products/ingres
\n
Oracle Database (Oracle Corporation)\n
An enterprise level database management system with a free to use Express Edition.
\n\n
http://www.oracle.com/technetwork/products/express-edition/overview/index.html
\n
IBM DB2 (IBM Corporation)\n
A powerful relational database management system with a free edition called DB2 Express-C.
\n\n
http://www-01.ibm.com/software/data/db2/express/
\n
PostgreSQL\n
Open Source relational database management system with tons of features.
\n\n
http://www.postgresql.org/
\n
MySQL (Oracle Corporation)\n
Popular and easy to use open source DBMS.
\n\n
http://www.mysql.com/
\n
Firebird\n
Full featured, open source relational DBMS.
\n\n
http://www.firebirdsql.org/
\n
SQLite (D. Richard Hipp)\n
Popular, small and free to use embeddable database system.
\n\n
http://sqlite.org/
\n
Access (Microsoft Corporation)\n
Personal relational database system with a graphical interface.
\n\n
http://office.microsoft.com/access
\n
SQL Server (Microsoft Corporation)\n
Proprietary, powerful dbms with a free to use express edition.
\n\n
https://www.microsoft.com/en-us/sqlserver/editions/2012-editions/express.aspx
\n

\n\n\n

Glossary

\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n\n

Alias	A temporary name given to a table in the FROM clause
Cross Join	A join listing all possible combination of rows without filtering
Database	A collection of organized data. Can be stored in a digital format like on a computer
DBMS	Database Management System. A software to control and manage digital databases
Field	A column in a table
Foreign Key	A column in a table that matches a primary key column in another table
Normalization	Breaking down a raw database into tables and removing redundancies
Record	A row of a table
SQL	Structured Query Language. A language used to interact with databases
Table	A matrix like display/abstraction of data in row-column format

\n\n

\n\n","tocStructure",[349,350,351,352,353,354,355,356,357,358,359,360,361,362,363,364,365,366,367,368,369,370,371,372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407,408,409,410,411,412,413,414,415,416,417,418,419,420],{"_74":75,"_42":76,"_68":-5,"_77":78,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":84,"_68":-5,"_77":85,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":88,"_68":-5,"_77":89,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":92,"_68":-5,"_77":93,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":97,"_68":-5,"_77":98,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":102,"_68":-5,"_77":103,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":105,"_68":-5,"_77":106,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":109,"_68":-5,"_77":110,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":112,"_68":-5,"_77":113,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":115,"_68":-5,"_77":116,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":118,"_68":-5,"_77":119,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":121,"_68":-5,"_77":122,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":124,"_68":-5,"_77":125,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":127,"_68":-5,"_77":128,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":131,"_68":-5,"_77":132,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":134,"_68":-5,"_77":135,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":137,"_68":-5,"_77":138,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":140,"_68":-5,"_77":141,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":143,"_68":-5,"_77":144,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":147,"_68":-5,"_77":148,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":150,"_68":-5,"_77":151,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":153,"_68":-5,"_77":154,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":156,"_68":-5,"_77":157,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":159,"_68":-5,"_77":160,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":162,"_68":-5,"_77":163,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":165,"_68":-5,"_77":166,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":169,"_68":-5,"_77":170,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":172,"_68":-5,"_77":173,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":175,"_68":-5,"_77":176,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":178,"_68":-5,"_77":179,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":181,"_68":-5,"_77":182,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":184,"_68":-5,"_77":185,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":188,"_68":-5,"_77":189,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":191,"_68":-5,"_77":192,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":194,"_68":-5,"_77":195,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":197,"_68":-5,"_77":198,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":200,"_68":-5,"_77":201,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":204,"_68":-5,"_77":205,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":207,"_68":-5,"_77":208,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":210,"_68":-5,"_77":211,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":213,"_68":-5,"_77":214,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":216,"_68":-5,"_77":217,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":220,"_68":-5,"_77":221,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":223,"_68":-5,"_77":224,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":226,"_68":-5,"_77":227,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":229,"_68":-5,"_77":230,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":232,"_68":-5,"_77":233,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":236,"_68":-5,"_77":237,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":239,"_68":-5,"_77":240,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":242,"_68":-5,"_77":243,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":245,"_68":-5,"_77":246,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":249,"_68":-5,"_77":250,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":252,"_68":-5,"_77":253,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":255,"_68":-5,"_77":256,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":258,"_68":-5,"_77":259,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":261,"_68":-5,"_77":262,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":264,"_68":-5,"_77":265,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":268,"_68":-5,"_77":269,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":271,"_68":-5,"_77":272,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":274,"_68":-5,"_77":275,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":277,"_68":-5,"_77":278,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":280,"_68":-5,"_77":281,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":283,"_68":-5,"_77":284,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":287,"_68":-5,"_77":288,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":290,"_68":-5,"_77":291,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":293,"_68":-5,"_77":294,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":297,"_68":-5,"_77":298,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":300,"_68":-5,"_77":301,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":96,"_42":303,"_68":-5,"_77":304,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":306,"_68":-5,"_77":307,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":310,"_68":-5,"_77":311,"_70":79,"_99":-5,"_100":-5,"_82":-5},{"_74":75,"_42":314,"_68":-5,"_77":315,"_70":79,"_99":-5,"_100":-5,"_82":-5},"publicChapterContent","nextChapter","sections",[],"routes/_store+/_layout","routes/_store+/_readers+/_layout","actionData","errors"]

About the author

Table of Contents

Preface

About the author

Acknowledgements

\n1. An Introduction to SQL

\n1.1 SQL Commands Classification

\n1.2 Explaining Tables

\n2. Getting your database ready

\n2.1 Using Ingres

\n2.2 Using SQLite

\n2.3 Creating your own database

\n2.4 Table Creation

\n2.5 Inserting data

\n2.6 Writing your first query

\n3. Constraints

\n3.1 Selective fields INSERT

\n3.2 Primary Key Constraint

\n3.3 Unique Key Constraint

\n3.4 Differences between a Primary Key and a Unique Key

\n4. Operations on Tables

\n4.1 Dropping Tables

\n4.2 Creating new tables from existing tables

\n4.3 Modifying tables

\n4.4 Verifying the result in Ingres

\n4.5 Verifying the result in other DBMS’s

\n4.6 Showing table information in SQLite

\n5. Writing Basic Queries

\n5.1 Selecting a limited number of columns

\n5.2 Ordering the results

\n5.3 Ordering using field abbreviations

\n5.4 Putting conditions with WHERE

\n5.5 Combining conditions

\n6. Manipulating Data

\n6.1 Inserting NULL’s

\n6.2 Inserting data into a table from another table

\n6.3 Updating existing data

\n6.4 Deleting data from tables

\n7. Organizing your data

\n7.1 Normalization

\n7.2 Atomicity

\n7.3 Repeating Groups

\n7.4 Splitting the table

\n8. Doing more with queries

\n8.1 Counting the records in a table

\n8.2 Column Aliases

\n8.3 Order of execution of SELECT queries

\n8.4 Using the LIKE operator

\n9. Calculated Fields

\n9.1 Mathematical calculations

\n9.2 String operations

\n9.3 Literal Values

\n10. Aggregation and Grouping

\n10.1 Aggregate Functions

\n10.2 Using DISTINCT with COUNT

\n10.3 Using MIN to find minimum values

\n10.4 Grouping Data

\n10.5 The HAVING Clause

\n11. Understanding Joins

\n11.1 What is a Join?

\n11.2 Alternative Join Syntax

\n11.3 Resolving ambiguity in join columns

\n11.4 Cross Joins

\n11.5 Self Joins

\n12. Subqueries

\n12.1 Types of subqueries

\n12.2 Using subqueries in INSERT statements

\n13. Working with Sets

\n13.1 Union

\n13.2 Intersection

\n13.3 Difference

Further Reading

Appendix: Major Database Management Systems

Glossary