Tutorial
Ian Shields | Published February 24, 2016
InfrastructureLinuxSystems
In this tutorial, learn about Structured Query Language (SQL), including:
This tutorial is a brief introduction to the SQL concepts that you need to know for the LPI 102 exam.
So far in this series of tutorials, you used flat text files to store data. Flat text files can be suitable for relatively small amounts of data, but they do not lend themselves well to storing large amounts of data or to querying that data. Over the years, several kinds of databases were developed for that purpose, including hierarchical and network databases, but the most common is now the relational database. The relational database is based on the ideas of E. F. Codd, who worked for IBM® and published the seminal paper “A Relational Model of Data for Large Shared Data Banks” in 1970. Several relational database systems exist today, including commercial products such as IBM DB2®, IBM Informix®, and Oracle Database, and open source projects such as MySQL, PostgreSQL SQLite, and MariaDB (a fork of MySQL). Relational databases use SQL as a data definition and query language.
This series of tutorials helps you learn Linux system administration tasks. You can also use the material in these tutorials to prepare for the Linux Professional Institute’s LPIC-1: Linux Server Professional Certification exams.
See ” Learn Linux, 101: A roadmap for LPIC-1” for a description of and link to each tutorial in this series. The roadmap is in progress and reflects the version 4.0 objectives of the LPIC-1 exams as updated April 15, 2015. As tutorials are completed, they will be added to the roadmap.
This tutorial helps you prepare for Objective 105.3 in Topic 105 of the Linux Server Professional (LPIC-1) exam 102. The objective has a weight of 2.
To get the most from the tutorials in this series, you need a basic knowledge of Linux and a working Linux system on which you can practice the commands covered in this tutorial. You should be familiar with GNU and UNIX® commands. Sometimes, different versions of a program format output differently, so your results might not always look exactly like the listings shown here.
For this tutorial, you need a database, such as MariaDB, and the documentation for your database.
The SQL examples in this tutorial are largely distribution and database independent. Most use the open source MariaDB version 10.0.21 on Fedora 23 with a 4.2.6 kernel. I also include a few examples that use IBM DB2 Express-C version 10.5, a no-charge community edition of the IBM DB2 data server, on CentOS 6.7 with a 2.6.32 kernel. By comparing these examples, you can get an idea of the kinds of differences that you might encounter when you work with multiple database systems. If you need to implement portable database programming or scripts, you need to learn more about the ISO/ANSI SQL standards than I can cover here.
Some database-manipulation commands are database-specific. Small differences in SQL syntax, particularly for nonstandard extensions, also exist. Consult the documentation for the database that you are using, as necessary.
A relational database consists of a set of tables. Think of each row of data in the table as a record, with each column of the table corresponding to the fields in the record for the corresponding row. The data in a column is all of the same type— such as character, integer, date, or binary data (such as images). By using structured data in this way, you can construct meaningful queries such as “find all the employees who were hired after a certain date” or “find all the parcels of land that are larger than 0.25 acres.”
The data in a relational database can have a unique value in each row, such as an employee ID for an employee record, a parcel ID for a municipal land database, or a user ID for a computer system. If so, you probably create an index for that column, and you probably designate one such column as your primary key for the table. Keys and indexes help the database engine to speed up retrieval of your data.
When you install a database, you usually install a database server. You might also install a client, or you can access the server through programmed applications or the command line. The server usually runs as a daemon process, and your client or application usually connects to it over a TCP/IP connection. You must take steps to secure the installation and prevent access that you do not want to support. Database installation and security are beyond the scope of the LPI 102 exam objectives.
Many Linux distributions come with packages for MariaDB, SQLite, PostgreSQL, and MySQL. For this tutorial, I primarily use MariaDB. See “Distributions Which Include MariaDB,” or use the appropriate package-management tool for your distribution to check what is already packaged for your distribution. MariaDB is an open source fork of MySQL, so mysql is part of many of the command names.
mysql
IBM DB2 Express-C, the community-supported version of IBM DB2, is “free to develop, free to deploy, free to distribute.” To use DB2 Express-C, download it (see Related topics), decompress it, and run the db2setup program. You can use the db2prereqcheck command from the package to check that your system meets the prerequisites. After installation, log in as the db2inst1 user that you created while running db2setup and run the db2sampl command to install the DB2 sample database that I use in this tutorial.
db2setup
db2prereqcheck
db2sampl
DB2 uses Linux user authentication to validate users. To use the db2inst1 instance as another user, copy the last few lines from the db2inst1 user’s .bashrc (~db2inst1/.bashrc) file to your own .bashrc. These lines run the db2profile script for the instance to set up environment variables and add the DB2 directories to your PATH variable.
PATH
I give you a few brief tips here to help you get running with MariaDB on a distribution that packages it. I use Fedora 23 in this tutorial. Start by installing the mariadb and mariadb-server packages, which pull in several other packages that are needed.
mariadb
mariadb-server
Next, run the mysql_secure_installation command as a user with root authority. By running this command, you:
mysql_secure_installation
The database root user is not the system root user and should not have the same password. You can rerun mysql_secure_installation if you need to make changes.
For this tutorial, I use the employee sample database that is available as part of the test_db package from GitHub. Listing 1 shows the steps that I used to install the database on my system.
Listing 1. Installing the sample employee database
[ian@attic-f23 ~]$ unzip -q test_db-master.zip [ian@attic-f23 ~]$ cd test_db-master [ian@attic-f23 test_db-master]$ mysql -u root -p <employees.sql Enter password: INFO CREATING DATABASE STRUCTURE INFO storage engine: InnoDB INFO LOADING departments INFO LOADING employees INFO LOADING dept_emp INFO LOADING dept_manager INFO LOADING titles INFO LOADING salaries data_load_time_diff NULL
The default InnoDB engine used in is fine for this tutorial’s purposes. See Related topics for more information on database engines.
Now that MariaDB is installed on your system, you can start to see what you have. Listing 2 uses the mysqlshow command to show the databases that I have installed. The -u option specifies the database root user, and the -p option tells mysqlshow to prompt you for the password that you defined when you ran the mysql_secure_installation command.
mysqlshow
-u
-p
Listing 2. What databases do I have?
[ian@attic-f23 ~]$ mysqlshow -u root -p Enter password: +--------------------+ | Databases | +--------------------+ | employees | | information_schema | | mysql | | performance_schema | | test | +--------------------+
You can see that I have five databases: the employees database that I just created, the test database that I did not delete, and three other databases. Database programs usually include several databases to describe the database itself, and you see them in Listing 2.
employees
test
The mysqlshow command is an convenience tool for quickly listing information about databases, tables, and columns. MariaDB (and MySQL) include an interactive command-line interface (CLI) program called mysql that is similar to a database shell. DB2 also has a CLI program, called db2. As with shells such as bash, you can pass a single command to either of these database shells, or you can run an interactive session with many commands. Listing 3 shows the database information via the mysql command with the -e option to execute a single database command.
db2
-e
Listing 3. Listing MariaDB databases
[ian@attic-f23 ~]$ mysql -u root -p -e "show databases" Enter password: +--------------------+ | Database | +--------------------+ | employees | | information_schema | | mysql | | performance_schema | | test | +--------------------+
If you also installed DB2 Express-C, you created a user named db2inst1 (the default) to manage the database. Listing 4 shows how to use the db2 command to get corresponding information about DB2 databases.
Listing 4. Listing DB2 databases
[ian@attic4-cent ~]$ db2 list database directory System Database Directory Number of entries in the directory = 1 Database 1 entry: Database alias = SAMPLE Database name = SAMPLE Local database directory = /home/db2inst1 Database release level = 10.00 Comment = Directory entry type = Indirect Catalog database partition number = 0 Alternate server hostname = Alternate server port number =
Before you can extract information from a database, you need to know what is in the database. In Listing 5, you see how to:
show tables
describe
Listing 5. Showing MariaDB table and column information
[ian@attic-f23 ~]$ mysql -u root -p employees Enter password: Reading table information for completion of table and column names You can turn off this feature to get a quicker startup with -A Welcome to the MariaDB monitor. Commands end with ; or \g. Your MariaDB connection id is 19 Server version: 10.0.21-MariaDB MariaDB Server Copyright (c) 2000, 2015, Oracle, MariaDB Corporation Ab and others. Type 'help;' or '\h' for help. Type '\c' to clear the current input statement. MariaDB [employees]> show tables; +----------------------+ | Tables_in_employees | +----------------------+ | current_dept_emp | | departments | | dept_emp | | dept_emp_latest_date | | dept_manager | | employees | | salaries | | titles | +----------------------+ 8 rows in set (0.00 sec) MariaDB [employees]> describe employees; +------------+---------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+---------------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | birth_date | date | NO | | NULL | | | first_name | varchar(14) | NO | | NULL | | | last_name | varchar(16) | NO | | NULL | | | gender | enum('M','F') | NO | | NULL | | | hire_date | date | NO | | NULL | | +------------+---------------+------+-----+---------+-------+ 6 rows in set (0.00 sec)
Each field (or column) in the employees table has several attributes:
int(11)
varchar(16)
PRI
auto_increment
If you need help in the shell, use the help (or ?) command. Listing 6 shows the help output for the describe command.
help
?
Listing 6. Help for the MariaDB describe command
MariaDB [employees]> ? describe Name: 'DESCRIBE' Description: Syntax: {DESCRIBE | DESC} tbl_name [col_name | wild] DESCRIBE provides information about the columns in a table. It is a shortcut for SHOW COLUMNS FROM. These statements also display information for views. (See [HELP SHOW COLUMNS].) col_name can be a column name, or a string containing the SQL "%" and "_" wildcard characters to obtain output only for the columns with names matching the string. There is no need to enclose the string within quotation marks unless it contains spaces or other special characters. MariaDB> DESCRIBE City; +------------+----------+------+-----+---------+----------------+ | Field | Type | Null | Key | Default | Extra | +------------+----------+------+-----+---------+----------------+ | Id | int(11) | NO | PRI | NULL | auto_increment | | Name | char(35) | NO | | | | | Country | char(3) | NO | UNI | | | | District | char(20) | YES | MUL | | | | Population | int(11) | NO | | 0 | | +------------+----------+------+-----+---------+----------------+ 5 rows in set (0.00 sec) The description for SHOW COLUMNS provides more information about the output columns (see [HELP SHOW COLUMNS]). URL: https://mariadb.com/kb/en/describe/
Notice that the help shows DESCRIBE in all-uppercase letters. In general, SQL command names are not case sensitive. Whether names for other objects — such as databases, tables, or columns — are case sensitive depends on your database program and the platform that it runs on. See Related topics for more information on names in MariaDB.
DESCRIBE
You saw examples of SQL data types in Show 4. Typical SQL data types fall broadly into four categories:
CHAR(8)
VARCHAR(240)
BINARY(12)
VARBINARY(500)
BLOB(200000)
SMALLINT
INT
INTEGER
BIGINT
FLOAT
DOUBLE
DECIMAL
TRUE
FALSE
DATE
TIME
These examples are not exhaustive, and different database programs might expand on them. For example, DB2 supports CLOB for character large objects and DBCLOB for large objects that contain double-byte character data. MariaDB supports TINYINT for a 1-byte integer and ENUM to store enumeration data, as you saw for the gender field in Listing 5.
CLOB
DBCLOB
TINYINT
ENUM
gender
Note: The use of enumeration data types in SQL is somewhat controversial. A more traditional approach uses a reference table with a foreign key. Search the Internet to learn more about the debate over enumeration data types in databases.
You have now seen examples of databases, tables, and the types of data in the tables. The next step is to get data out of the tables. You retrieve data by using the SELECT statement. You can select everything in the table, or you can select from particular columns. Listing 7 shows what is in the departments table, how to select everything from it via SELECT `SELECT`, and then how to select only the department names.
SELECT
departments
Listing 7. Basic use of the SELECT command
MariaDB [employees]> describe departments; +-----------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-----------+-------------+------+-----+---------+-------+ | dept_no | char(4) | NO | PRI | NULL | | | dept_name | varchar(40) | NO | UNI | NULL | | +-----------+-------------+------+-----+---------+-------+ 2 rows in set (0.00 sec) MariaDB [employees]> select * from departments; +---------+--------------------+ | dept_no | dept_name | +---------+--------------------+ | d009 | Customer Service | | d005 | Development | | d002 | Finance | | d003 | Human Resources | | d001 | Marketing | | d004 | Production | | d006 | Quality Management | | d008 | Research | | d007 | Sales | +---------+--------------------+ 9 rows in set (0.00 sec) MariaDB [employees]> select dept_name from departments; +--------------------+ | dept_name | +--------------------+ | Customer Service | | Development | | Finance | | Human Resources | | Marketing | | Production | | Quality Management | | Research | | Sales | +--------------------+ 9 rows in set (0.00 sec)
When you select everything from the departments table as in Listing 7, the output is sorted by the department name. If you want to sort your output data according to the values in one or more columns, you use the ORDER BYORDERBY clause. Specify a comma-separated list of columns. You can also specify ASC for ascending sort (the default) or DESC for descending sort order. Listing 8 shows how to sort your output by dept_no in descending order.
ORDERBY
ASC
DESC
dept_no
Listing 8. Sorting query output
MariaDB [employees]> select dept_name,dept_no from departments order by dept_no desc; +--------------------+---------+ | dept_name | dept_no | +--------------------+---------+ | Customer Service | d009 | | Research | d008 | | Sales | d007 | | Quality Management | d006 | | Development | d005 | | Production | d004 | | Human Resources | d003 | | Finance | d002 | | Marketing | d001 | +--------------------+---------+ 9 rows in set (0.00 sec)
You can choose which data to display by using the WHERE clause. You can compare columns, or compare a column value against a single value. Such a single value — called a scalar— can be a constant (which must be enclosed in quotation marks) or the result of a scalar function, such as today’s date. Use Boolean AND, OR, and NOT operators to define more-complex conditions. Listing 9 shows some examples, using the departments table.
WHERE
AND
OR
NOT
Listing 9. Using the WHERE clause to limit data output
MariaDB [employees]> select * from departments where dept_no > 'd007'; +---------+------------------+ | dept_no | dept_name | +---------+------------------+ | d008 | Research | | d009 | Customer Service | +---------+------------------+ 2 rows in set (0.00 sec) MariaDB [employees]> select * from departments -> WHERE dept_name = 'Customer Service' OR ( -> dept_no > 'd002' AND dept_no <= "d006" -> ); +---------+--------------------+ | dept_no | dept_name | +---------+--------------------+ | d009 | Customer Service | | d005 | Development | | d003 | Human Resources | | d004 | Production | | d006 | Quality Management | +---------+--------------------+ 5 rows in set (0.00 sec)
Notice in Listing 9 that the output from the more-complex WHERE clause is not sorted. Use the ORDER clause for your desired sort. Notice also that you can enter SQL commands on multiple lines. MariaDB gives you a -> prompt on the second and subsequent lines. I remove these secondary prompts from subsequent examples so that you can copy and paste the command directly into your own system more easily.
ORDER
->
You can also do pattern matching and direct comparison by using LIKE. Most SQL databases support the following two wildcard characters:
LIKE
%
_
Listing 10 shows how to find all department names that contain the string es.
es
Listing 10. Using the WHERE clause with LIKE for pattern matching
MariaDB [employees]> select * from departments WHERE dept_name LIKE '%es%'; +---------+-----------------+ | dept_no | dept_name | +---------+-----------------+ | d003 | Human Resources | | d008 | Research | | d007 | Sales | +---------+-----------------+ 3 rows in set (0.00 sec)
Some databases, including MariaDB, support regular-expression pattern matching. See the documentation for your database for more information on this topic.
You have seen some simple comparison operators. SQL also has several functions that you can use, including:
COUNT
DISTINCT
MAX
MIN
NOW
DATEDIFF
DAY
LEAST
Many other such functions are available, so consult your documentation to learn about them.
Listing 11 shows you how to count the number of rows in two different tables.
Listing 11. Using COUNT to count the number of rows in a table
MariaDB [employees]> # How many employees? MariaDB [employees]> select count(*) from employees; +----------+ | count(*) | +----------+ | 300024 | +----------+ 1 row in set (0.08 sec) MariaDB [employees]> # How many managers? MariaDB [employees]> select count(*) from dept_manager; +----------+ | count(*) | +----------+ | 24 | +----------+ 1 row in set (0.00 sec)
Listing 12 shows how to use the DISTINCT function to see if some managers were hired on the same date as others.
MariaDB [employees]> describe dept_manager; +-----------+---------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-----------+---------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | dept_no | char(4) | NO | PRI | NULL | | | from_date | date | NO | | NULL | | | to_date | date | NO | | NULL | | +-----------+---------+------+-----+---------+-------+ 4 rows in set (0.00 sec) MariaDB [employees]> select DISTINCT(from_date) from dept_manager order by from_date; +------------+ | from_date | +------------+ | 1985-01-01 | | 1988-09-09 | | 1988-10-17 | | 1989-05-06 | | 1989-12-17 | | 1991-03-07 | | 1991-04-08 | | 1991-09-12 | | 1991-10-01 | | 1992-03-21 | | 1992-04-25 | | 1992-08-02 | | 1992-09-08 | | 1994-06-28 | | 1996-01-03 | | 1996-08-30 | +------------+ 16 rows in set (0.00 sec)
Only 16 rows are returned. You could have combined COUNT and DISTINCT to find this number, but you also have the actual start dates in Listing 12. You don’t yet have any idea which date or dates have multiple managers starting.
Suppose you want to find out how long a manager has been managing. You can calculate the difference between the from_date and to_date, but how does this work for someone who is currently managing? Tables often use a marker of some kind to indicate a current date — possibly a NULL value, or possibly a date far in the future. The dept_manager table uses a future date for this purpose. Someone whose to_date is later than today’s date is still managing. Use the NOW function to get the current time stamp, or use the CURDATE function to get the current date only. Listing 13 shows one way to find how many days each manager has been a manager.
from_date
to_date
NULL
dept_manager
CURDATE
Listing 13. Finding how long managers have been managing
MariaDB [employees]> # Show marker for managers still managing MariaDB [employees]> select max(to_date) from dept_manager; +--------------+ | max(to_date) | +--------------+ | 9999-01-01 | +--------------+ 1 row in set (0.00 sec) MariaDB [employees]> # Calculate duration of management in days MariaDB [employees]> select emp_no, datediff(least(to_date, curdate()),from_date) from dept_manager; +--------+-----------------------------------------------+ | emp_no | datediff(least(to_date, curdate()),from_date) | +--------+-----------------------------------------------+ | 110022 | 2464 | | 110039 | 8869 | | 110085 | 1811 | | 110114 | 9522 | | 110183 | 2636 | | 110228 | 8697 | | 110303 | 1347 | | 110344 | 1423 | | 110386 | 1489 | | 110420 | 7074 | | 110511 | 2671 | | 110567 | 8662 | | 110725 | 1586 | | 110765 | 859 | | 110800 | 1020 | | 110854 | 7868 | | 111035 | 2256 | | 111133 | 9077 | | 111400 | 2288 | | 111534 | 9045 | | 111692 | 1385 | | 111784 | 1422 | | 111877 | 1212 | | 111939 | 7314 | +--------+-----------------------------------------------+ 24 rows in set (0.00 sec)
When you have a long expression such as datediff(least(to_date,curdate()),from_date), you might want to shorten the column heading or provide an alias for the expression because you plan to use it elsewhere in your query — for example, in a WHERE clause. Listing 14 shows how to use aliases to provide shorter headings, to find out how many years all the current managers have been managing.
datediff(least(to_date,curdate()),from_date)
Listing 14. Using aliases for shorter headings
MariaDB [employees]> select emp_no AS Employee, (datediff(least(to_date, curdate()),from_date)/365.25) as Years from dept_manager where to_date > curdate(); +----------+---------+ | Employee | Years | +----------+---------+ | 110039 | 24.2820 | | 110114 | 26.0698 | | 110228 | 23.8111 | | 110420 | 19.3676 | | 110567 | 23.7153 | | 110854 | 21.5414 | | 111133 | 24.8515 | | 111534 | 24.7639 | | 111939 | 20.0246 | +----------+---------+ 9 rows in set (0.00 sec)
GROUP BY
Sometimes you want aggregate information from a table. For example, you want to know how many employees are in each $10K salary range. You can use the GROUP BY clause to group your data for this purpose. Listing 15 shows how.
Listing 15. Aggregating data by using GROUP BY
MariaDB [employees]> describe salaries; +-----------+---------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-----------+---------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | salary | int(11) | NO | | NULL | | | from_date | date | NO | PRI | NULL | | | to_date | date | NO | | NULL | | +-----------+---------+------+-----+---------+-------+ 4 rows in set (0.00 sec) MariaDB [employees]> select salary DIV 10000 as 10K, count(*) as Number from salaries where to_date > curdate() group by 10K; +------+--------+ | 10K | Number | +------+--------+ | 3 | 85 | | 4 | 20220 | | 5 | 44666 | | 6 | 56236 | | 7 | 49128 | | 8 | 32351 | | 9 | 19939 | | 10 | 10611 | | 11 | 4729 | | 12 | 1645 | | 13 | 421 | | 14 | 78 | | 15 | 15 | +------+--------+ 13 rows in set (1.42 sec)
HAVING
In Group 1, the values in the Number column are computed from aggregate data. Suppose that you are only interested in salary ranges that have 5,000 or fewer employees in the range. Your first idea might be to use a WHERE clause, but you cannot use it for intermediate result data that is computed as an aggregate. You need to use a HAVING clause to restrict the intermediate results to a subset that has a particular condition or combination of conditions. Listing 16 shows how to find salary ranges that have 5,000 or fewer employees in the range.
Number
Listing 16. Using the HAVING clause
MariaDB [employees]> select salary DIV 10000 as 10K, count(*) as Number from salaries where to_date > curdate() group by 10K HAVING Number <= 5000; +------+--------+ | 10K | Number | +------+--------+ | 3 | 85 | | 11 | 4729 | | 12 | 1645 | | 13 | 421 | | 14 | 78 | | 15 | 15 | +------+--------+ 6 rows in set (1.45 sec)
JOIN
So far, the examples in this tutorial each use a single table. The salaries table and the dept_manager table contain the employee number but not the name or other employee information. Employee information is kept in the employees table. Keeping data in separate tables is how SQL design eliminates (or at least reduces) both redundancy in data storage and the associated risk of data being updated in one place but not another.
salaries
When you want to extract information such as the name and gender of all current managers, you need to get this information from both the dept_manager table and the employees table. For this purpose, you use the JOIN clause with a conditional expression that specifies the join condition. Most commonly, you join two tables on a single field, where the field value is the same in both tables — for example, the emp_no field that occurs in both the dept_manager table and the employees table.
emp_no
Listing 17 shows how to use JOIN to find the name and gender of all current managers. (Note the use of the aliases e and m for the dept_manager table and the employees table.)
e
m
Listing 17. Using the JOIN clause for current manager information
MariaDB [employees]> SELECT e.first_name, e.last_name, e.gender FROM employees as e JOIN dept_manager as m ON e.emp_no = m.emp_no WHERE m.to_date > now() ORDER BY e.last_name; +------------+------------+--------+ | first_name | last_name | gender | +------------+------------+--------+ | Leon | DasSarma | F | | Oscar | Ghazalie | M | | Hilary | Kambil | F | | Isamu | Legleitner | F | | Vishwani | Minakawa | M | | Dung | Pesch | M | | Karsten | Sigstam | F | | Yuchang | Weedman | M | | Hauke | Zhang | M | +------------+------------+--------+ 9 rows in set (0.00 sec)
If you want salary instead of gender as an output column, you must join the salaries table to the other two tables, as shown in Listing 18. (I added some parentheses, which I often find helpful in constructing complex queries.)
Listing 18. Using the JOIN clause for current manager salaries
MariaDB [employees]> SELECT e.first_name, e.last_name, s.salary FROM (employees as e JOIN dept_manager as m ON e.emp_no = m.emp_no) JOIN salaries as s on e.emp_no = s.emp_no WHERE m.to_date > now() AND s.to_date > now() ORDER BY e.last_name; +------------+------------+--------+ | first_name | last_name | salary | +------------+------------+--------+ | Leon | DasSarma | 74510 | | Oscar | Ghazalie | 56654 | | Hilary | Kambil | 79393 | | Isamu | Legleitner | 83457 | | Vishwani | Minakawa | 106491 | | Dung | Pesch | 72876 | | Karsten | Sigstam | 65400 | | Yuchang | Weedman | 58745 | | Hauke | Zhang | 101987 | +------------+------------+--------+ 9 rows in set (0.00 sec)
The JOIN examples in Join 1 and Join 2 are called inner joins: They find the rows from both tables that match the join condition. The three other common kinds of joins are:
LEFTJOIN
RIGHTJOIN
OUTERJOIN
FULLOUTERJOIN
The INNER JOININNERJOIN is the most common and is the default if no join type is specified. Listing 19 illustrates a LEFT JOINLEFTJOIN to show selected employees and their from_date and to_date if they happen to be a manager. In this example, I also use a LIMIT value to limit the output to a maximum of 15 rows.
INNERJOIN
LIMIT
Listing 19. Using the LEFT JOIN clause
MariaDB [employees]> SELECT e.emp_no, e.first_name, e.last_name, s.salary, m.from_date, m.to_date FROM (employees as e LEFT JOIN dept_manager as m ON e.emp_no = m.emp_no) JOIN salaries as s on e.emp_no = s.emp_no AND s.to_date > now() WHERE e.last_name LIKE 'Kambi%' AND e.first_name > 'G' ORDER BY e.last_name, e.first_name limit 15; +--------+------------+-----------+--------+------------+------------+ | emp_no | first_name | last_name | salary | from_date | to_date | +--------+------------+-----------+--------+------------+------------+ | 431582 | Gaurav | Kambil | 118128 | NULL | NULL | | 252478 | Gaurav | Kambil | 69516 | NULL | NULL | | 487991 | Gift | Kambil | 115960 | NULL | NULL | | 204311 | Gil | Kambil | 96756 | NULL | NULL | | 416604 | Gonzalo | Kambil | 80009 | NULL | NULL | | 236164 | Hausi | Kambil | 66130 | NULL | NULL | | 412003 | Hausi | Kambil | 83213 | NULL | NULL | | 111534 | Hilary | Kambil | 79393 | 1991-04-08 | 9999-01-01 | | 295702 | Huei | Kambil | 49498 | NULL | NULL | | 77408 | Idoia | Kambil | 67122 | NULL | NULL | | 271049 | Jianhao | Kambil | 58393 | NULL | NULL | | 216820 | JiYoung | Kambil | 87541 | NULL | NULL | | 206261 | Jongsuk | Kambil | 78396 | NULL | NULL | | 250164 | Josyula | Kambil | 98835 | NULL | NULL | | 289558 | Jouko | Kambil | 51393 | NULL | NULL | +--------+------------+-----------+--------+------------+------------+ 15 rows in set (0.14 sec)
Sometimes you are interested in only part of the data in a query, and you want to manipulate just that part. For this purpose, you can use a subselect (also called a subquery), which is essentially a SELECT within another SELECT. You can also use a subselect in a FROM, WHERE, or HAVING clause. Sometimes the work done by a subselect can also be done by a JOIN. With so many possibilities, I show you two examples to get you started.
FROM
The titles table in the employees database shows which positions an employee has held. Some employees have held multiple positions. Listing 20 shows how to find the positions held by three employees. Note the use of the IN operator to select the results from a set and the BETWEEN operator to select employee numbers between two values.
titles
IN
BETWEEN
Listing 20. Employees having held different positions
MariaDB [employees]> describe titles; +-----------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +-----------+-------------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | title | varchar(50) | NO | PRI | NULL | | | from_date | date | NO | PRI | NULL | | | to_date | date | YES | | NULL | | +-----------+-------------+------+-----+---------+-------+ 4 rows in set (0.00 sec) MariaDB [employees]> select * from titles where emp_no IN (10001, 10004, 499666); +--------+--------------------+------------+------------+ | emp_no | title | from_date | to_date | +--------+--------------------+------------+------------+ | 10001 | Senior Engineer | 1986-06-26 | 9999-01-01 | | 10004 | Engineer | 1986-12-01 | 1995-12-01 | | 10004 | Senior Engineer | 1995-12-01 | 9999-01-01 | | 499666 | Assistant Engineer | 1987-10-18 | 1994-10-18 | | 499666 | Engineer | 1994-10-18 | 2001-10-18 | | 499666 | Senior Engineer | 2001-10-18 | 9999-01-01 | +--------+--------------------+------------+------------+ 6 rows in set (0.00 sec) MariaDB [employees]> select * from titles where emp_no BETWEEN 10001 AND 10004; +--------+-----------------+------------+------------+ | emp_no | title | from_date | to_date | +--------+-----------------+------------+------------+ | 10001 | Senior Engineer | 1986-06-26 | 9999-01-01 | | 10002 | Staff | 1996-08-03 | 9999-01-01 | | 10003 | Senior Engineer | 1995-12-03 | 9999-01-01 | | 10004 | Engineer | 1986-12-01 | 1995-12-01 | | 10004 | Senior Engineer | 1995-12-01 | 9999-01-01 | +--------+-----------------+------------+------------+ 5 rows in set (0.01 sec)
You already know how to use GROUP BYGROUPBY and COUNT to find out how many positions each employee has held. What if you only want to know how many employees have held a single position, how many have held two positions, and so on? Listing 21 shows how you construct a query to find the number of positions held by each employee, then how to use this query as a subselect in a query to answer the question of how many employees have held a single position, how many have held two positions, and so on. I name the subquery subq. The subquery returns only a single column: jobs. The main query refers to this column by using the qualified name subq.jobs.
GROUPBY
subq
jobs
subq.jobs
Listing 21. Using a subselect to find out how many employees have held multiple positions
MariaDB [employees]> SELECT count(*) as jobs from titles group by emp_no limit 5; +------+ | jobs | +------+ | 1 | | 1 | | 1 | | 2 | | 2 | +------+ 5 rows in set (0.00 sec) MariaDB [employees]> SELECT jobs as '# jobs', count(subq.jobs) as '# employees' FROM ( SELECT count(*) as jobs from titles group by emp_no ) subq group by jobs; +--------+-------------+ | # jobs | # employees | +--------+-------------+ | 1 | 159754 | | 2 | 137256 | | 3 | 3014 | +--------+-------------+ 3 rows in set (0.31 sec)
Now suppose that you would like to know who the top 10 salary earners are in the company. Listing 22 shows another subselect that returns this information.
Listing 22. Using a subselect to find top salary earners
MariaDB [employees]> SELECT e.emp_no, e.last_name, e.first_name, ( SELECT MAX(salary) FROM salaries s WHERE e.emp_no = s.emp_no ) max_sal FROM employees e ORDER BY max_sal desc limit 10; +--------+-----------+------------+---------+ | emp_no | last_name | first_name | max_sal | +--------+-----------+------------+---------+ | 43624 | Pesch | Tokuyasu | 158220 | | 254466 | Mukaidono | Honesty | 156286 | | 47978 | Whitcomb | Xiahua | 155709 | | 253939 | Luders | Sanjai | 155513 | | 109334 | Alameldin | Tsutomu | 155377 | | 80823 | Baca | Willard | 154459 | | 493158 | Meriste | Lidong | 154376 | | 205000 | Griswold | Charmane | 153715 | | 266526 | Chenoweth | Weijing | 152710 | | 237542 | Hatcliff | Weicheng | 152687 | +--------+-----------+------------+---------+ 10 rows in set (2.66 sec)
I mentioned that you can sometimes do the work of a subselect by using JOIN. Listing 23 shows how you can find the top 10 salary earners by using a JOIN. Note that this query executes faster than the subselect of Listing 22. In general, prefer a join over a subselect if performance matters. The performance difference depends on your database engine.
Listing 23. Using a JOIN to find top salary earners
MariaDB [employees]> SELECT e.emp_no, e.last_name, e.first_name, max(s.salary) as max_sal FROM employees e JOIN salaries s ON e.emp_no = s.emp_no GROUP BY e.emp_no ORDER BY max_sal desc limit 10; +--------+-----------+------------+---------+ | emp_no | last_name | first_name | max_sal | +--------+-----------+------------+---------+ | 43624 | Pesch | Tokuyasu | 158220 | | 254466 | Mukaidono | Honesty | 156286 | | 47978 | Whitcomb | Xiahua | 155709 | | 253939 | Luders | Sanjai | 155513 | | 109334 | Alameldin | Tsutomu | 155377 | | 80823 | Baca | Willard | 154459 | | 493158 | Meriste | Lidong | 154376 | | 205000 | Griswold | Charmane | 153715 | | 266526 | Chenoweth | Weijing | 152710 | | 237542 | Hatcliff | Weicheng | 152687 | +--------+-----------+------------+---------+ 10 rows in set (2.16 sec)
In the “” section, I mentioned that enumeration types in SQL are somewhat controversial. The gender column in the employees table is an ENUM with values M and F. These values look like characters but are actually stored internally as integers. They are sorted in an ORDER BYORDERBY clause according to their internal numeric representation, which can give a possibly surprising result. Listing 24 shows the first five employees sorted by gender.
M
F
Listing 24. Ordering output via ENUM fields
MariaDB [employees]> SELECT * FROM ( SELECT * FROM employees LIMIT 5 ) x ORDER BY x.gender; +--------+------------+------------+-----------+--------+------------+ | emp_no | birth_date | first_name | last_name | gender | hire_date | +--------+------------+------------+-----------+--------+------------+ | 10001 | 1953-09-02 | Georgi | Facello | M | 1986-06-26 | | 10003 | 1959-12-03 | Parto | Bamford | M | 1986-08-28 | | 10004 | 1954-05-01 | Chirstian | Koblick | M | 1986-12-01 | | 10005 | 1955-01-21 | Kyoichi | Maliniak | M | 1989-09-12 | | 10002 | 1964-06-02 | Bezalel | Simmel | F | 1985-11-21 | +--------+------------+------------+-----------+--------+------------+ 5 rows in set (0.00 sec)
It might surprise you to see the M values listed before the F values. If you want the alphabetic collating order, use a CAST or CONVERT in the ORDER BYORDERBY clause. For example:
CAST
CONVERT
ORDER BY CAST(x.gender AS CHAR)
You learned about tables and how to explore their structure. You also learned how to create SQL queries to find answers from the data in one or more tables. In the remainder of this tutorial, I show you how to create and manipulate tables and how to insert, update, and delete the data in your tables.
Suppose that you decide to create a reference table for employee gender, rather than use an enumeration. You decide to key off the initial of the gender, which is M or F in English. You also want a second column for the full word, MALE or FEMALE. The first step is to create the table by using CREATE TABLECREATETABLE. The basic form of CREATE TABLECREATETABLE provides a table name followed by a list of columns, indexes, and constraints. Listing 25 shows how to create the gender table with two columns, a single character as the primary key, and a variable character field of up to 20 characters.
MALE
FEMALE
CREATETABLE
Listing 25. Creating a new table
MariaDB [employees]> CREATE TABLE gender ( code CHAR(1) NOT NULL, gender VARCHAR(20) NOT NULL, PRIMARY KEY (code) ); Query OK, 0 rows affected (0.23 sec) MariaDB [employees]> describe gender; +--------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +--------+-------------+------+-----+---------+-------+ | code | char(1) | NO | PRI | NULL | | | gender | varchar(20) | NO | | NULL | | +--------+-------------+------+-----+---------+-------+ 2 rows in set (0.00 sec)
Now that you have an empty table, you need to populate it with data. Use the INSERT command to load one or more rows of data. You specify a list of columns that you want to insert data into, and you then specify one or more lists of matching values. Listing 26 shows how to add the two rows that you planned.
INSERT
Listing 26. Inserting data in the gender table
MariaDB [employees]> INSERT INTO gender (code, gender) VALUES ('F', 'FEMALE'), ('M', 'MALE'); Query OK, 2 rows affected (0.03 sec) Records: 2 Duplicates: 0 Warnings: 0 MariaDB [employees]> select * from gender; +------+--------+ | code | gender | +------+--------+ | F | FEMALE | | M | MALE | +------+--------+ 2 rows in set (0.00 sec)
Assume next that you think some employees might need to be added to the database before you know their gender, so you decide to add a third possibility for UNASSIGNED gender. Listing 27 shows another way to insert data: by using the SET clause.
UNASSIGNED
SET
Listing 27. Inserting a new row in the gender table
MariaDB [employees]> INSERT INTO gender SET code='U', gender='UNSIGNED'; Query OK, 1 row affected (0.04 sec) MariaDB [employees]> select * from gender; +------+----------+ | code | gender | +------+----------+ | F | FEMALE | | M | MALE | | U | UNSIGNED | +------+----------+ 3 rows in set (0.00 sec)
The sample database I am using has only two gender values. Suppose that your company adopts a gender diversity policy and you need additional values for gender diverse employees. Use either the VALUES option or SET option with INSERT INTO to insert more rows into your table.
VALUES
INSERT INTO
Because of a typographical error, I inserted UNSIGNED when I meant to insert UNASSIGNED. Use the UPDATE command to fix this mistake, as shown in Listing 28.
UNSIGNED
UPDATE
Listing 28. Updating a table value
MariaDB [employees]> UPDATE gender SET gender='UNASSIGNED' WHERE code='U'; Query OK, 1 row affected (0.03 sec) Rows matched: 1 Changed: 1 Warnings: 0 MariaDB [employees]> select * from gender; +------+------------+ | code | gender | +------+------------+ | F | FEMALE | | M | MALE | | U | UNASSIGNED | +------+------------+ 3 rows in set (0.00 sec)
Now the human resources manager tells you that the company always knows a new employee’s gender before the employee is added to the database, so you need to delete the entry for unassigned gender. Use the DELETE command, as shown in Listing 29.
DELETE
Listing 29. Deleting a table row
MariaDB [employees]> DELETE FROM gender WHERE code='U'; Query OK, 1 row affected (0.04 sec) MariaDB [employees]> select * from gender; +------+--------+ | code | gender | +------+--------+ | F | FEMALE | | M | MALE | +------+--------+ 2 rows in set (0.00 sec)
Sometimes you want to change a table. Some of the many things you can do include:
Listing 30 shows how to alter the gender table to add a new column, gender_fr for a French version of the gender column, load the new French data, then rename the existing gender column to gender_en.
gender_fr
gender_en
Listing 30. Altering the gender table
MariaDB [employees]> ALTER TABLE gender ADD COLUMN gender_fr VARCHAR(20) AFTER gender; Query OK, 0 rows affected (0.38 sec) Records: 0 Duplicates: 0 Warnings: 0 MariaDB [employees]> UPDATE gender SET gender_fr='MASCULIN' WHERE code='M'; Query OK, 1 row affected (0.03 sec) Rows matched: 1 Changed: 1 Warnings: 0 MariaDB [employees]> UPDATE gender SET gender_fr='FMININ' WHERE code='F'; Query OK, 1 row affected (0.04 sec) Rows matched: 1 Changed: 1 Warnings: 0 MariaDB [employees]> ALTER TABLE gender CHANGE COLUMN gender gender_en VARCHAR(20); Query OK, 0 rows affected (0.38 sec) Records: 0 Duplicates: 0 Warnings: 0 MariaDB [employees]> select * from gender; +------+-----------+-----------+ | code | gender_en | gender_fr | +------+-----------+-----------+ | F | FEMALE | FÉMININ | | M | MALE | MASCULIN | +------+-----------+-----------+ 2 rows in set (0.00 sec)
You can also create a table and populate it with data from a SELECT. Listing 31 shows a simple example of how to create a table of female employees.
Listing 31. Creating a table of female employees
MariaDB [employees]> create table female_employees select * from employees where gender='F'; Query OK, 120051 rows affected (3.22 sec) Records: 120051 Duplicates: 0 Warnings: 0
When you create a table this way, the new table probably will not inherit all the attributes of the source table. Compare the table descriptions in Listing 32. Notice that the female_employees table has no key.
female_employees
Listing 32. Comparing the employees and female_employees tables
MariaDB [employees]> describe employees; +------------+---------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+---------------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | birth_date | date | NO | | NULL | | | first_name | varchar(14) | NO | | NULL | | | last_name | varchar(16) | NO | | NULL | | | gender | enum('M','F') | NO | | NULL | | | hire_date | date | NO | | NULL | | +------------+---------------+------+-----+---------+-------+ 6 rows in set (0.00 sec) MariaDB [employees]> describe female_employees; +------------+---------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+---------------+------+-----+---------+-------+ | emp_no | int(11) | NO | | NULL | | | birth_date | date | NO | | NULL | | | first_name | varchar(14) | NO | | NULL | | | last_name | varchar(16) | NO | | NULL | | | gender | enum('M','F') | NO | | NULL | | | hire_date | date | NO | | NULL | | +------------+---------------+------+-----+---------+-------+ 6 rows in set (0.00 sec)
Something else to consider if you create a table from existing data is how you plan to keep the data in the two tables synchronized in the future. In this case, you probably want a VIEW, which enables you to use the result of the SELECT as if it were a table. The data stays in the underlying table or tables, and only a single copy needs to be updated. Listing 33 shows how to create a view for male employees. Note the word AS between the view name and the SELECT.
VIEW
AS
Listing 33. Creating a view for male employees
MariaDB [employees]> CREATE VIEW male_employees AS select * from employees where gender='M'; Query OK, 0 rows affected (0.05 sec) MariaDB [employees]> select * from male_employees limit 5; +--------+------------+------------+-----------+--------+------------+ | emp_no | birth_date | first_name | last_name | gender | hire_date | +--------+------------+------------+-----------+--------+------------+ | 10001 | 1953-09-02 | Georgi | Facello | M | 1986-06-26 | | 10003 | 1959-12-03 | Parto | Bamford | M | 1986-08-28 | | 10004 | 1954-05-01 | Chirstian | Koblick | M | 1986-12-01 | | 10005 | 1955-01-21 | Kyoichi | Maliniak | M | 1989-09-12 | | 10008 | 1958-02-19 | Saniya | Kalloufi | M | 1994-09-15 | +--------+------------+------------+-----------+--------+------------+ 5 rows in set (0.00 sec)
If you want to know which tables in a database are really views, a command is available to help you. Different databases use different commands. In MariaDB use:
show full tables
When you create a table, you can combine your column specifications and the result of a SELECT. For the final example of table creation, Listing 34 shows you one way to make a copy of the employees table with a CHAR(1) field for gender instead of the enumeration and how to set this value by using an IF statement.
CHAR(1)
IF
Listing 34. Making a revised copy of the employees table
MariaDB [employees]> CREATE TABLE employees_new ( emp_no int(11) NOT NULL, birth_date date NOT NULL, first_name varchar(14) NOT NULL, last_name varchar(16) NOT NULL, gender CHAR(1), hire_date date NOT NULL, PRIMARY KEY (emp_no) ) select emp_no, birth_date, first_name, last_name, IF(gender = 'M', 'M', 'F') as gender, hire_date from employees; Query OK, 300024 rows affected (5.92 sec) Records: 300024 Duplicates: 0 Warnings: 0 MariaDB [employees]> describe employees_new; +------------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+-------------+------+-----+---------+-------+ | emp_no | int(11) | NO | PRI | NULL | | | birth_date | date | NO | | NULL | | | first_name | varchar(14) | NO | | NULL | | | last_name | varchar(16) | NO | | NULL | | | gender | char(1) | YES | | NULL | | | hire_date | date | NO | | NULL | | +------------+-------------+------+-----+---------+-------+ 6 rows in set (0.00 sec)
The creation and maintenance of foreign keys is beyond the scope of this introductory tutorial. However, you can still use a JOIN to join the employees_new and gender table to perform queries such as those that I illustrate in Listing 35.
employees_new
Listing 35. Joining the employees_new and gender tables
MariaDB [employees]> select e.emp_no, e.first_name, e.last_name, e.gender, g.gender_en, g.gender_fr from employees_new e join gender g on e.gender=g.code limit 5; +--------+------------+-----------+--------+-----------+-----------+ | emp_no | first_name | last_name | gender | gender_en | gender_fr | +--------+------------+-----------+--------+-----------+-----------+ | 10001 | Georgi | Facello | M | MALE | MASCULIN | | 10002 | Bezalel | Simmel | F | FEMALE | FÉMININ | | 10003 | Parto | Bamford | M | MALE | MASCULIN | | 10004 | Chirstian | Koblick | M | MALE | MASCULIN | | 10005 | Kyoichi | Maliniak | M | MALE | MASCULIN | +--------+------------+-----------+--------+-----------+-----------+ 5 rows in set (0.00 sec)
I showed you how to use DELETE to delete rows from a table. If you want to drop a column, rather than a row, you need to use ALTER TABLEALTERTABLE. For example, if you want to delete the gender column from the female_employees table, use the command illustrated in Listing 36.
ALTERTABLE
Listing 36. Deleting the gender column from the female_employees table
MariaDB [employees]> ALTER TABLE female_employees DROP COLUMN gender; Query OK, 0 rows affected (4.32 sec) Records: 0 Duplicates: 0 Warnings: 0 MariaDB [employees]> describe female_employees; +------------+-------------+------+-----+---------+-------+ | Field | Type | Null | Key | Default | Extra | +------------+-------------+------+-----+---------+-------+ | emp_no | int(11) | NO | | NULL | | | birth_date | date | NO | | NULL | | | first_name | varchar(14) | NO | | NULL | | | last_name | varchar(16) | NO | | NULL | | | hire_date | date | NO | | NULL | | +------------+-------------+------+-----+---------+-------+ 5 rows in set (0.00 sec)
To drop the female_employees table and male_employees view, use the DROP TABLEDROPTABLE and DROP VIEWDROPVIEW commands shown in Listing 37.
male_employees
DROPTABLE
DROPVIEW
Listing 37. Deleting the female_employees table and male_employees view
MariaDB [employees]> drop table female_employees; Query OK, 0 rows affected (0.14 sec) MariaDB [employees]> drop view male_employees; Query OK, 0 rows affected (0.00 sec)
If you want to remove the entire employees database and all of its tables, use the DROP DATABASEDROPDATABASE command, as shown in Listing 38.
DROPDATABASE
Listing 38. Deleting the entire employees database
MariaDB [employees]> drop database employees; Query OK, 10 rows affected (0.92 sec)
This concludes your brief introduction to SQL. I have only scratched the surface of a large subject, so I encourage you to use this tutorial as a starting point for further exploration.
Get the Code »
IBM Cloud PrivateIBM LinuxONE+
This article provides the resources to facilitate standard operating systems and network protocols on IBM's leadership supercomputing hardware.
IBM Power SystemsInfrastructure+
QEMU provides support for virtual machines to use SCSI storage directly with SCSI pass-through, using the virtio-blk or virtio-scsi storage…
DatabasesInfrastructure+
Back to top