DB2 9 Fundamentals exam 730 prep, Part 4:

Working with DB2 data

Skill Level: Introductory

Roman Melnyk (roman_b_melnyk@hotmail.com)

Staff Information Development
IBM Toronto

24 Jul 2006

This tutorial introduces you to Structured Query Language (SQL) and helps to give
you a good understanding of how DB2® 9 uses SQL to manipulate data in a
relational database. This tutorial is the fourth in a series of seven tutorials that you
can use to help prepare for DB2 9 Fundamentals Certification (Exam 730).

Section 1. Before you start

About this series

Thinking about seeking certification on DB2 fundamentals (Exam 730)? If so, you've
landed in the right spot. This series of seven DB2 certification preparation tutorials
covers all the basics -- the topics you'll need to understand before you read the first
exam question. Even if you're not planning to seek certification right away, this set of
tutorials is a great place to begin learning what's new in DB2 9.

About this tutorial

This tutorial explains how DB2 uses SQL to manipulate data in a relational
database.. The material provided herein primarily covers the objectives in Section 4
of the exam, which is entitled "Working with DB2 Data using SQL and XQuery." You
can view these objectives at: http://www-03.ibm.com/certify/tests/obj730.shtml.

Topics covered in this tutorial include:

• An introduction to SQL

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 1 of 22
developerWorks® ibm.com/developerWorks

• A description of Data Manipulation Language (DML) and examples that

demonstrate how to use it to obtain specific information.
• Transaction boundaries
• SQL procedures and user-defined functions

Objectives
After completing this tutorial, you should be able to:

• Understand the fundamentals of SQL, with a focus on SQL language

elements
• Use DML to select, insert, update, or delete data
• Use COMMIT and ROLLBACK statements to manage transactions, and
know what constitutes a transaction boundary
• Create and call SQL procedures or invoke user-defined functions from the
command line

System Requirements
If you haven't already done so, you can download the free trial version of IBM DB2 9
to work along with this tutorial. Installing DB2 helps you understand many of the
concepts that are tested on the DB2 9 Fundamentals Certification exam. DB2
installation is not covered in this tutorial, but the process is documented in the DB2
Information Center.

Conventions used in this tutorial

The following text highlighting conventions are used in this tutorial:

• Monospaced text is used for SQL statements. UPPERCASE text identifies

SQL keywords, and lowercase text identifies user-supplied values in
example code.
• Except in the code, database object names are given in uppercase
characters, and table column names are given in mixed case.
• Italic text is used when introducing a new term, or to identify a parameter
variable.
All of the examples in this tutorial are based on the SAMPLE database, which comes
with DB2. Because sample output is provided in most cases, you do not need
access to the product to understand the examples.

Working with DB2 data

Page 2 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

Section 2. Structured Query Language (SQL)

The parts of speech of SQL

SQL is a language that is used to define and manipulate database objects. Use SQL
to define a database table, insert data into the table, change the data in the table,
and retrieve data from the table. Like all languages, SQL has a defined syntax and a
set of language elements.

Most SQL statements contain one or more of the following language elements:

• Single-byte characters can be a letter (A-Z, a-z, $, #, and @, or a member

of an extended character set), a digit (0-9), or a special character
(including the comma, asterisk, plus sign, percent sign, ampersand, and
several others).
• A token is a sequence of one or more characters. It cannot contain blank
characters unless it is a delimited identifier (one or more characters
enclosed by double quotation marks) or a string constant.
• An SQL identifier is a token that is used to form a name.
• The data type of a value determines how DB2 interprets that value. DB2
supports many built-in data types and also supports user-defined types
(UDTs).
• A constant specifies a value. They are classified as character, graphic, or
hexadecimal string constants, or integer, decimal, or floating-point
numeric constants.
• A special register is a storage area that is defined for an application
process by the database manager and is used to store information that
can be referenced in SQL statements. Examples of special registers are
CURRENT DATE, CURRENT DBPARTITIONNUM, and CURRENT SCHEMA.
• A routine can be a function, a method, or a procedure.
• A function is a relationship between one or more input data values
and one or more result values. Database functions can be either
built-in or user-defined.
Column (or aggregate) functions operate on a set of values in a
column to return a single value. For example:

• SUM(sales) returns the sum of the values in the Sales

column.
• AVG(sales) returns the sum of the values in the Sales

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 3 of 22
developerWorks® ibm.com/developerWorks

column divided by the number of values in that column.

• MIN(sales) returns the smallest value in the Sales column.
• MAX(sales) returns the largest value in the Sales column.
• COUNT(sales) returns the number of non-null values in the
Sales column.
Scalar functions operate on a single value to return another single
value. For example:

• ABS(-5) returns the absolute value of -5 -- that is, 5.

• HEX(69) returns the hexadecimal representation of the
number 69 -- that is, 45000000.
• LENGTH('Pierre') returns the number of bytes in the string
"Pierre" -- that is, 6. For a GRAPHIC string, the LENGTH
function returns the number of double-byte characters.
• YEAR('03/14/2002') extracts the year portion of
03/14/2002 -- that is, 2002.
• MONTH('03/14/2002') extracts the month portion of
03/14/2002 -- that is, 3.
• DAY('03/14/2002') extracts the day portion of 03/14/2002
-- that is, 14.
• LCASE('SHAMAN') or LOWER('SHAMAN') returns a string in
which all of the characters have been converted to lowercase
characters -- that is, 'shaman'.
• UCASE('shaman') or UPPER('shaman') returns a string in
which all of the characters have been converted to uppercase
characters -- that is, 'SHAMAN'.
User-defined functions are registered to a database in the system
catalog (accessible through the SYSCAT.ROUTINES catalog view)
using the CREATE FUNCTION statement.

• A method is also a relationship between a set of input data values

and a set of result values. However, database methods are defined,
either implicitly or explicitly, as part of the definition of a user-defined
structured type. For example, a method called CITY (of type
ADDRESS) is passed an input value of type VARCHAR, and the result
is a subtype of ADDRESS. User-defined methods are registered to a
database in the system catalog (accessible through the
SYSCAT.ROUTINES catalog view) using the CREATE METHOD
statement. For more information about structured types, refer to An
introduction to structured data types and typed tables .
• A procedure is an application program that can be started by
executing a CALL statement. The arguments of a procedure are

Working with DB2 data

Page 4 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

individual scalar values that can be of different types and that can be
used to pass values into the procedure, receive return values from the
procedure, or both. User-defined procedures are registered to a
database in the system catalog (accessible through the
SYSCAT.ROUTINES catalog view) using the CREATE PROCEDURE
statement.
• An expression specifies a value. There are string expressions, arithmetic
expressions, and case expressions, which can be used to specify a
particular result based on the evaluation of one or more conditions.
• A predicate specifies a condition that is true, false, or unknown about a
given row or group. There are subtypes:
• A basic predicate compares two values (for example, x > y).
• The BETWEEN predicate compares a value with a range of values.
• The EXISTS predicate tests for the existence of certain rows.
• The IN predicate compares one or more values with a collection of
values.
• The LIKE predicate searches for strings that have a certain pattern.
• The NULL predicate tests for null values.

Section 3. Data Manipulation Language (DML)

Using the SELECT statement to retrieve data from database

tables
The SELECT statement is used to retrieve table or view data. In its simplest form, the
SELECT statement can be used to retrieve all the data in a table. For example, to
retrieve all the STAFF data from the SAMPLE database, issue the following
command:

SELECT * FROM staff

Here is a partial result set returned by this query:

ID NAME DEPT JOB YEARS SALARY COMM

10 Sanders 20 Mgr 7 18357.50 -
20 Pernal 20 Sales 8 18171.25 612.45
30 Marenghi 38 Mgr 5 17506.75 -

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 5 of 22
developerWorks® ibm.com/developerWorks

To restrict the number of rows in a result set, use the FETCH FIRST clause. For
example:

SELECT * FROM staff FETCH FIRST 10 ROWS ONLY

Retrieve specific columns from a table by specifying a select list of column names
separated by commas. For example:

SELECT name, salary FROM staff

Use the DISTINCT clause to eliminate duplicate rows in a result set. For example:

SELECT DISTINCT dept, job FROM staff

Use the AS clause to assign a meaningful name to an expression or an item in the

select list. For example:

SELECT name, salary + comm AS pay FROM staff

Without the AS clause, the derived column would have been named 2, indicating that
it is the second column in the result set.

Using the WHERE clause and predicates to limit the amount of

data returned by a query
Use the WHERE clause to select specific rows from a table or view by specifying one
or more selection criteria, or search conditions. A search condition consists of one or
more predicates. A predicate specifies something about a row that is either true or
false (see The parts of speech of SQL). When building search conditions, be sure to:

• Apply arithmetic operations only to numeric data types

• Make comparisons only among compatible data types
• Enclose character values within single quotation marks
• Specify character values exactly as they appear in the database
Let's look at some examples.

• Find the names of staff members whose salaries are greater than
$20,000:

"SELECT name, salary FROM staff

WHERE salary > 20000"

Working with DB2 data

Page 6 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

Enclosing the statement within double quotation marks keeps your

operating system from misinterpreting special characters, such as * or >;
the greater-than symbol could be interpreted as an output redirection
request.
• List the name, job title, and salary of staff members who are not
managers and whose salary is greater than $20,000:

"SELECT name, job, salary FROM staff

WHERE job <> 'Mgr'
AND salary > 20000"

• Find all names that start with the letter S:

SELECT name FROM staff

WHERE name LIKE 'S%'

In this example, the percent sign (%) is a wild card character that
represents a string of zero or more characters.
A subquery is a SELECT statement that appears within the WHERE clause of a main
query and feeds its result set to that WHERE clause. For example:

"SELECT lastname FROM employee

WHERE lastname IN
(SELECT sales_person FROM sales
WHERE sales_date < '01/01/1996')"

A correlation name is defined in the FROM clause of a query and can serve as a
convenient short name for a table. Correlation names also eliminate ambiguous
references to identical column names from different tables. For example:

"SELECT e.salary FROM employee e

WHERE e.salary <
(SELECT AVG(s.salary) FROM staff s)"

Using the ORDER BY clause to sort results

Use the ORDER BY clause to sort the result set by values in one or more columns.
The column names that are specified in the ORDER BY clause do not have to be
specified in the select list. For example:

"SELECT name, salary FROM staff

WHERE salary > 20000
ORDER BY salary"

Sort the result set in descending order by specifying DESC in the ORDER BY clause:

ORDER BY salary DESC

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 7 of 22
developerWorks® ibm.com/developerWorks

Using joins to retrieve data from more than one table

A join is a query that combines data from two or more tables. It is often necessary to
select information from two or more tables because required data is often
distributed. A join adds columns to the result set. For example, a full join of two
three-column tables produces a result set with six columns.

The simplest join is one in which there are no specified conditions. For example:

SELECT deptnumb, deptname, manager, id, name, dept, job

FROM org, staff

This statement returns all combinations of rows from the ORG table and the STAFF
table. The first three columns come from the ORG tables, and the last four columns
come from the STAFF table. Such a result set (the cross product of the two tables) is
not very useful. What is needed is a join condition to refine the result set. For
example, here is a query that is designed to identify staff members who are
managers:

SELECT deptnumb, deptname, id AS manager_id, name AS manager

FROM org, staff
WHERE manager = id
ORDER BY deptnumb

And here is a partial result set returned by this query:

DEPTNUMB DEPTNAME MANAGER_ID MANAGER

10 Head Office 160 Molinare
15 New England 50 Hanes
20 Mid Atlantic 10 Sanders

The statement you looked at in the last section is an example of an inner join. Inner
joins return only rows from the cross product that meet the join condition. If a row
exists in one table but not the other, it is not included in the result set. To explicitly
specify an inner join, rewrite the previous query with an INNER JOIN operator in the
FROM clause:

...
FROM org INNER JOIN staff
ON manager = id
...

The keyword ON specifies the join conditions for the tables being joined. DeptNumb
and DeptName are columns in the ORG table, and Manager_ID and Manager are
based on columns (ID and Name) in the STAFF table. The result set for the inner

Working with DB2 data

Page 8 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

join consists of rows that have matching values for the Manager and ID columns in
the left table (ORG) and the right table (STAFF), respectively. (When you perform a
join on two tables, you arbitrarily designate one table to be the left table and the
other to be the right.)

Outer joins return rows that are generated by an inner join operation, plus rows that
would not be returned by the inner join operation. There are three types of outer
joins:

• A left outer join includes the inner join plus the rows from the left table that
are not returned by the inner join. This type of join uses the LEFT OUTER
JOIN (or LEFT JOIN) operator in the FROM clause.
• A right outer join includes the inner join plus the rows from the right table
that are not returned by the inner join. This type of join uses the RIGHT
OUTER JOIN (or RIGHT JOIN) operator in the FROM clause.
• A full outer join includes the inner join plus the rows from both the left
table and the right table that are not returned by the inner join. This type
of join uses the FULL OUTER JOIN (or FULL JOIN) operator in the
FROM clause.
Construct more complex queries to answer more difficult questions. The following
query is designed to generate a list of employees who are responsible for projects,
identifying those employees who are also managers by listing the departments that
they manage:

SELECT empno, deptname, projname

FROM (employee
LEFT OUTER JOIN project
ON respemp = empno)
LEFT OUTER JOIN department
ON mgrno = empno

The first outer join gets the name of any project for which the employee is
responsible; this outer join is enclosed by parentheses and is resolved first. The
second outer join gets the name of the employee's department if that employee is a
manager.

Using set operators to combine two or more queries into a

single query
Combine two or more queries into a single query by using the UNION, EXCEPT, or
INTERSECT set operators. Set operators process the results of the queries,
eliminate duplicates, and return the final result set.

• The UNION set operator generates a result table by combining two or

more other result tables.
• The EXCEPT set operator generates a result table by including all rows

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 9 of 22
developerWorks® ibm.com/developerWorks

that are returned by the first query, but not by the second or any
subsequent queries.
• The INTERSECT set operator generates a result table by including only
rows that are returned by all the queries.
Following is an example of a query that makes use of the UNION set operator. The
same query could use the EXCEPT or the INTERSECT set operator by substituting
the appropriate keyword for UNION.

"SELECT sales_person FROM sales

WHERE region = 'Ontario-South'
UNION
SELECT sales_person FROM sales
WHERE sales > 3"

Using the GROUP BY clause to summarize results

Use the GROUP BY clause to organize rows in a result set. Each group is
represented by a single row in the result set. For example:

SELECT sales_date, MAX(sales) AS max_sales FROM sales

GROUP BY sales_date

This statement returns a list of sales dates from the SALES table. The SALES table
in the SAMPLE database contains sales data, including the number of successful
transactions by a particular sales person on a particular date. There is typically more
than one record per date. The GROUP BY clause groups the data by sales date, and
the MAX function in this example returns the maximum number of sales recorded for
each sales date.

A different flavor of the GROUP BY clause includes the specification of the

GROUPING SETS clause. Grouping sets can be used to analyze data at different
levels of aggregation in a single pass. For example:

SELECT YEAR(sales_date) AS year, region, SUM(sales) AS tot_sales

FROM sales
GROUP BY GROUPING SETS (YEAR(sales_date), region, () )

Here, the YEAR function is used to return the year portion of date values, and the
SUM function is used to return the total in each set of grouped sales figures. The
grouping sets list specifies how the data is to be grouped, or aggregated. A pair of
empty parentheses is added to the grouping sets list to get a grand total in the result
set. The statement returns the following:

YEAR REGION TOT_SALES

- - 155
- Manitoba 41

Working with DB2 data

Page 10 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

- Ontario-North 9
- Ontario-South 52
- Quebec 53
1995 - 8
1996 - 147

A statement that is almost identical to the previous one, but that specifies the
ROLLUP clause, or the CUBE clause instead of the GROUPING SETS clause, returns
a result set that provides a more detailed perspective on the data. It might provide
summaries by location or time.

The HAVING clause is often used with a GROUP BY clause to retrieve results for
groups that satisfy only a specific condition. A HAVING clause can contain one or
more predicates that compare some property of the group with another property of
the group or a constant. For example:

"SELECT sales_person, SUM(sales) AS total_sales FROM sales

GROUP BY sales_person
HAVING SUM(sales) > 25"

This statement returns a list of salespeople whose sales totals exceed 25.

Using the INSERT statement to add new rows to tables or

views
The INSERT statement is used to add new rows to a table or a view. Inserting a row
into a view also inserts the row into the table on which the view is based.

• Use a VALUES clause to specify column data for one or more rows. For
example:

INSERT INTO staff VALUES (1212,'Cerny',20,'Sales',3,90000.00,30000.00)

INSERT INTO staff VALUES (1213,'Wolfrum',20,'Sales',2,90000.00,10000.00)

Or the equivalent:

INSERT INTO staff (id, name, dept, job, years, salary, comm)
VALUES
(1212,'Cerny',20,'Sales',3,90000.00,30000.00),
(1213,'Wolfrum',20,'Sales',2,90000.00,10000.00)

• Specify a fullselect to identify data that is to be copied from other tables or

views. A fullselect is a statement that generates a result table. For
example:

CREATE TABLE pers LIKE staff

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 11 of 22
developerWorks® ibm.com/developerWorks

INSERT INTO pers

SELECT id, name, dept, job, years, salary, comm
FROM staff
WHERE dept = 38

Using the UPDATE statement to change data in tables or views

The UPDATE statement is used to change the data in a table or a view. Change the
value of one or more columns for each row that satisfies the conditions specified by
a WHERE clause. For example:

UPDATE staff
SET dept = 51, salary = 70000
WHERE id = 750

Or the equivalent:

UPDATE staff
SET (dept, salary) = (51, 70000)
WHERE id = 750

If you don't specify a WHERE clause, DB2 updates each row in the table or view!

Using the DELETE statement to get rid of data

The DELETE statement is used to delete entire rows of data from a table. Delete
each row that satisfies the conditions specified by a WHERE clause. For example:

DELETE FROM staff

WHERE id IN (1212, 1213)

If you don't specify a WHERE clause, DB2 deletes all the rows in the table or view!

Using the MERGE statement to combine conditional update,

insert, or delete operations
The MERGE statement updates a target table or updatable view using data from a
source table. During a single operation, rows in the target that match the source can
be updated or deleted, and rows that do not exist in the target can be inserted.

For example, consider the EMPLOYEE table to be the target table that contains
up-to-date information about the employees of a large company. Branch offices
handle updates to local employee records by maintaining their own version of the
EMPLOYEE table called MY_EMP. The MERGE statement can be used to update the
EMPLOYEE table with information that is contained in a MY_EMP table, which is the

Working with DB2 data

Page 12 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

source table for the merge operation.

The following statement inserts a row for new employee number 000015 into the
MY_EMP table.

INSERT INTO my_emp (empno, firstnme, midinit, lastname, workdept,

phoneno, hiredate, job, edlevel, sex, birthdate, salary)
VALUES ('000015', 'MARIO', 'M', 'MALFA', 'A00',
'6669', '05/05/2000', 'ANALYST', 15, 'M', '04/02/1973', 59000.00)

And the following statement inserts updated salary data for existing employee
number 000010 into the MY_EMP table.

INSERT INTO my_emp (empno, firstnme, midinit, lastname, edlevel, salary)

VALUES ('000010', 'CHRISTINE', 'I', 'HAAS', 18, 66600.00)

At this point, the inserted data exists only in the MY_EMP table because it has not
yet been merged with the EMPLOYEE table. Following is the MERGE statement that
takes the contents of the MY_EMP table and integrates them with the EMPLOYEE
table.

MERGE INTO employee AS e

USING (SELECT
empno, firstnme, midinit, lastname, workdept, phoneno,
hiredate, job, edlevel, sex, birthdate, salary
FROM my_emp) AS m
ON e.empno = m.empno
WHEN MATCHED THEN
UPDATE SET (salary) = (m.salary)
WHEN NOT MATCHED THEN
INSERT (empno, firstnme, midinit, lastname, workdept, phoneno,
hiredate, job, edlevel, sex, birthdate, salary)
VALUES (m.empno, m.firstnme, m.midinit, m.lastname,
m.workdept, m.phoneno, m.hiredate, m.job, m.edlevel,
m.sex, m.birthdate, m.salary)

Correlation names have been assigned to both the source and the target table to
avoid ambiguous table references in the search condition. The statement identifies
the columns in the MY_EMP table that are to be considered. The statement also
specifies the actions that are to be taken when a row in the MY_EMP table is found
to have a match in the EMPLOYEE table, or when a row does not have a match.

The following query executed against the EMPLOYEE table now returns a record for
employee 000015:

SELECT * FROM employee WHERE empno = '000015'

And the following query returns the record for employee 000010 with an updated
value for the SALARY column.

SELECT * FROM employee WHERE empno = '000010'

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 13 of 22
developerWorks® ibm.com/developerWorks

Using the data-change-table-reference clause to retrieve

intermediate result sets in the same unit of work
Suppose you want to give employee 000220 a 7% raise and retrieve her old salary,
both in the same unit of work (UOW). You could accomplish this by using a
data-change-table-reference clause, which is part of the FROM clause in an
SQL statement.

SELECT salary FROM OLD TABLE (

UPDATE employee SET salary = salary * 1.07
WHERE empno = '000220'
);
SALARY
-----------
29840.00
1 record(s) selected.

Columns in the target of the data-change operation (insert, update, or delete)

become the columns of the intermediate result table, and can be referenced by
name (in this case, Salary) in the select list of the query. The keywords OLD TABLE
specify that the intermediate result table is to contain values that precede the
data-change operation. The keywords NEW TABLE specify that the intermediate
result table is to contain values that immediately follow the data-change operation
(before referential integrity evaluation and the firing of after triggers has taken place).
The keywords FINAL TABLE specify that the intermediate result table is to contain
values that follow the data-change operation, referential integrity evaluation, and the
firing of after triggers.

Suppose you have a CUSTOMERS table that is defined as follows:

CREATE TABLE customers (

cust_id INTEGER GENERATED ALWAYS AS IDENTITY (
START WITH 10001
),
cust_name VARCHAR(12),
PRIMARY KEY (cust_id)
);

The primary key for this table, Cust_ID, is an automatically generated identity
column. You can use a data-change-table-reference clause to retrieve the
generated identity column value that is being used as a customer number.

SELECT * FROM FINAL TABLE (

INSERT INTO customers (cust_name) VALUES ('Lamarr')
);
CUST_ID CUST_NAME
----------- ------------
10001 Lamarr
1 record(s) selected.

Working with DB2 data

Page 14 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

Section 4. The COMMIT and ROLLBACK statements

and transaction boundaries

Units of work and savepoints

A unit of work (UOW), also known as a transaction, is a recoverable sequence of
operations within an application process. The classic example of a UOW is a simple
bank transaction to transfer funds from one account to another. There is an
inconsistency -- immediately after the application subtracts an amount of money
from one account -- that is resolved after an equal amount of money is added to the
second account. When these changes have been committed, they become available
to other applications.

A UOW starts implicitly when the first SQL statement within an application process is
issued against the database. All subsequent reads and writes by the same
application process are considered part of the same UOW. The application ends the
UOW by issuing either a COMMIT or a ROLLBACK statement, whichever is
appropriate. The COMMIT statement makes all changes made within the UOW
permanent, whereas the ROLLBACK statement reverses those changes. If the
application ends normally without an explicit COMMIT or ROLLBACK statement, the
UOW is automatically committed. If the application ends abnormally before the end
of a UOW, that unit of work is automatically rolled back.

A savepoint lets you selectively roll back a subset of actions that make up a UOW
without losing the entire transaction. You can nest savepoints and have several
savepoint levels active at the same time; this allows your application to roll back to a
specific savepoint, as necessary. Suppose you have three savepoints (A, B, and C)
defined within a particular UOW:

do some work;
savepoint A;
do some more work;
savepoint B;
do even more work;
savepoint C;
wrap it up;
roll back to savepoint B;

The rollback to savepoint B automatically releases savepoint C, but savepoints A

and B remain active.

For more information about savepoint levels and a detailed example illustrating
DB2's savepoint support, see Resources.

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 15 of 22
developerWorks® ibm.com/developerWorks

Section 5. SQL procedures and user-defined functions

Creating and calling an SQL procedure

An SQL procedure is a stored procedure whose body is written in SQL. The body
contains the logic of the SQL procedure. It can include variable declarations,
condition handling, flow-of-control statements, and DML. Multiple SQL statements
can be specified within a compound statement, which groups several statements
together into an executable block.

An SQL procedure is created when you successfully invoke a CREATE PROCEDURE

(SQL) statement, which defines the SQL procedure with an application server. SQL
procedures are a handy way to define more complex queries or tasks that can be
called whenever they are needed.

An easy way to create an SQL procedure is to code the CREATE PROCEDURE

(SQL) statement in a command line processor (CLP) script. For example, if the
statement shown below were in a file called createSQLproc.db2, that file could
be executed to create the SQL procedure:

1. Connect to the SAMPLE database.

2. Issue the following command:

db2 -td@ -vf createSQLproc.db2

This db2 command specifies the -td option flag, which tells the command line
processor to define and to use @ as the statement termination character (because
the semicolon is already being used as a statement termination character inside the
procedure body); the -v option flag, which tells the command line processor to echo
command text to standard output; and the -f option flag, which tells the command
line processor to read command input from the specified file instead of from
standard input.

CREATE PROCEDURE sales_status

(IN quota INTEGER, OUT sql_state CHAR(5))
DYNAMIC RESULT SETS 1
LANGUAGE SQL
BEGIN
DECLARE SQLSTATE CHAR(5);
DECLARE rs CURSOR WITH RETURN FOR
SELECT sales_person, SUM(sales) AS total_sales
FROM sales
GROUP BY sales_person
HAVING SUM(sales) > quota;
OPEN rs;
SET sql_state = SQLSTATE;
END @

Working with DB2 data

Page 16 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

This procedure, called SALES_STATUS, accepts an input parameter called quota

and returns an output parameter called sql_state. The procedure body consists of a
single SELECT statement that returns the name and the total sales figures for each
salesperson whose total sales exceed the specified quota.

Most SQL procedures accept at least one input parameter. In our example, the input
parameter contains a value (quota) that is used in the SELECT statement contained
in the procedure body.

Many SQL procedures return at least one output parameter. Our example includes
an output parameter (sql_state) that is used to report the success or failure of the
SQL procedure. DB2 returns an SQLSTATE value in response to conditions that
could be the result of an SQL statement. Because the returned SQLCODE or
SQLSTATE value pertains to the last SQL statement issued in the procedure body,
and accessing the values alters the subsequent values of these variables (because
an SQL statement is used to access them), the SQLCODE or SQLSTATE value should
be assigned to and returned through a locally defined variable (such as the sql_state
variable in our example).

The parameter list for an SQL procedure can specify zero or more parameters, each
of which can be one of three possible types:
• IN parameters pass an input value to an SQL procedure; this value
cannot be modified within the body of the procedure.
• OUT parameters return an output value from an SQL procedure.
• INOUT parameters pass an input value to an SQL procedure and return
an output value from the procedure.

SQL procedures can return zero or more result sets. In our example, the
SALES_STATUS procedure returns one result set. This has been done by:

1. Declaring the number of result sets that the SQL procedure returns in the
DYNAMIC RESULT SETS clause.

2. Declaring a cursor in the procedure body (using the WITH RETURN FOR
clause) for each result set that is returned. A cursor is a named control
structure that is used by an application program to point to a specific row
within an ordered set of rows. A cursor is used to retrieve rows from a set.

3. Opening the cursor for each result set that is returned.

4. Leaving the cursor(s) open when the SQL procedure returns.

Variables must be declared at the beginning of the SQL procedure body. To declare
a variable, assign a unique identifier to and specify an SQL data type for the variable
and, optionally, assign an initial value to the variable.

The SET clause in our sample SQL procedure is an example of a flow-of-control

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 17 of 22
developerWorks® ibm.com/developerWorks

clause. The following flow-of-control statements, structures, and clauses can be

used for conditional processing within an SQL procedure body:
• The CASE structure selects an execution path based on the evaluation of
one or more conditions.
• The FOR structure executes a block of code for each row of a table.
• The GET DIAGNOSTICS statement returns information about the
previous SQL statement into an SQL variable.
• The GOTO statement transfers control to a labeled block (a section of one
or more statements identified by a unique SQL name followed by a
colon).
• The IF structure selects an execution path based on the evaluation of
conditions. The ELSEIF and ELSE clauses enable you to branch or to
specify a default action if the other conditions are false.
• The ITERATE clause passes the flow of control to the beginning of a
labeled loop.
• The LEAVE clause transfers program control out of a loop or block of
code.
• The LOOP clause executes a block of code multiple times until a LEAVE,
ITERATE, or GOTO statement transfers control outside of the loop.
• The REPEAT clause executes a block of code until a specified search
condition returns true.
• The RETURN clause returns control from the SQL procedure to the caller.
• The SET clause assigns a value to an output parameter or SQL variable.
• The WHILE clause repeatedly executes a block of code while a specified
condition is true.

To successfully create SQL procedures, you must have installed the DB2
Application Development Client on the database server. (See the first tutorial in this
series for more on the Application Development Client.) The dependency on a C
compiler to create SQL procedures was eliminated in DB2 Universal Database
Version 8. All of the operations that were dependent on a C compiler are now
performed by DB2-generated byte code that is hosted in a virtual machine. For more
information about this enhancement, see Resources.

Use the SQL CALL statement to call SQL procedures from the DB2 command line.
The procedure being called must be defined in the system catalog. Client
applications written in any supported language can call SQL procedures. To call the
SQL procedure SALES_STATUS, perform the following steps:

1. Connect to the SAMPLE database.

2. Issue the following statement:

Working with DB2 data

Page 18 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

db2 CALL sales_status (25, ?)

Because parentheses have special meaning to the command shell on

UNIX-based systems, on those systems they must be preceded with a
backslash (\) character or be enclosed by double quotation marks:

db2 "CALL sales_status (25, ?)"

Do not include double quotation marks if you are using the command line
processor (CLP) in interactive input mode, characterized by the db2 =>
input prompt.

In this example, a value of 25 for the input parameter quota is passed to the SQL
procedure, as well as a question mark (?) place-holder for the output parameter
sql_state. The procedure returns the name and the total sales figures for each
salesperson whose total sales exceed the specified quota (25). The following is
sample output returned by this statement:

SQL_STATE: 00000

SALES_PERSON TOTAL_SALES
GOUNOT 50
LEE 91
"SALES_STATUS" RETURN_STATUS: "0"

Creating and using SQL user-defined functions

Create user-defined functions to extend the set of built-in DB2 functions. For
example, create functions that evaluate complex mathematical expressions or
manipulate strings, and then reference these functions in SQL statements like you
would any existing built-in function.

Suppose that you needed a function that returns the area of a circle when the radius
of that circle is specified as an argument to the function. Such a function is not
available as a built-in DB2 function, but you could create a user-defined SQL scalar
function to perform this task and reference the function wherever scalar functions
are supported within an SQL statement.

CREATE FUNCTION ca (r DOUBLE)

RETURNS DOUBLE
LANGUAGE SQL
CONTAINS SQL
NO EXTERNAL ACTION
DETERMINISTIC
RETURN 3.14159 * (r * r);

The NO EXTERNAL ACTION clause specifies that the function does not take any

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 19 of 22
developerWorks® ibm.com/developerWorks

action that changes the state of an object that the database manager does not
manage. The DETERMINISTIC keyword specifies that the function always returns
the same result for a given argument value. This information is used during query
optimization. A convenient way to execute the function is to reference it in a query.
In the following example, the query is executed (arbitrarily) against the
SYSIBM.SYSDUMMY1 catalog view, which only has one row:

db2 SELECT ca(96.8) AS area FROM sysibm.sysdummy1

AREA
------------------------
+2.94374522816000E+004
1 record(s) selected.

You can also create a user-defined table function, which takes zero or more input
arguments and returns data as a table. A table function can only be used in the
FROM clause of an SQL statement.

Suppose that you needed a function that returns the names and employee numbers
of all employees that hold a specific job, with the title of that job specified as an
argument to the function. The following is an example of a table function that
performs this task:

CREATE FUNCTION jobemployees (job VARCHAR(8))

RETURNS TABLE (
empno CHAR(6),
firstname VARCHAR(12),
lastname VARCHAR(15)
)
LANGUAGE SQL
READS SQL DATA
NO EXTERNAL ACTION
DETERMINISTIC
RETURN
SELECT empno, firstnme, lastname
FROM employee
WHERE employee.job = jobemployees.job;

The following query references the new table function in the FROM clause and
passes the job title 'CLERK' as the argument to the function. A correlation name,
introduced by the keyword AS, is required by the syntax:

db2 SELECT * FROM TABLE(jobemployees('CLERK')) AS clerk

EMPNO FIRSTNAME LASTNAME
------ ------------ ---------------
000120 SEAN O'CONNELL
000230 JAMES JEFFERSON
000240 SALVATORE MARINO
000250 DANIEL SMITH
000260 SYBIL JOHNSON
000270 MARIA PEREZ
6 record(s) selected.

Working with DB2 data

Page 20 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.
ibm.com/developerWorks developerWorks®

Section 6. Summary
This tutorial was designed to introduce you to Structured Query Language (SQL)
and to some of the ways that DB2 9 uses SQL to manipulate data in a relational
database. It also covered the fundamentals of SQL, including SQL language
elements, Data Manipulation Language (DML), SQL procedures, and user-defined
functions. Part 5: Working with DB2 objects, which discusses data types, tables,
views, and indexes as defined by DB2, will help you understand how to create and
use them.

To keep an eye on this series, bookmark the series page, DB2 9 DBA exam 730
prep tutorials.

Working with DB2 data

© Copyright IBM Corporation 1994, 2006. All rights reserved. Page 21 of 22
developerWorks® ibm.com/developerWorks

Resources
Learn
• Check out the other parts of the DB2 9 Fundamentals exam 730 prep tutorial
series.
• Certification exam site. Click the exam number to see more information about
Exams 730 and 731.
• DB2 9 overview. Find information about the new data server that includes
patented pureXML technology.
• DB2 XML evaluation guide: A step-by-step introduction to the XML storage and
query capabilities of DB2 9
• Find additional information about DB2 9 and SQL from the DB2 Information
Center.
• An IBM DB2 Universal Database "Stinger" Feature Preview: Enhanced
Savepoints offers more information on savepoints.
• What's New for SQL PL in the IBM DB2 Universal Database "Stinger" Release
explains the DB2 features that reduce the dependency on a C compiler to
create SQL procedures.
• Visit the developerWorks DBA Central zone to read articles and tutorials and
connect to other resources to expand your database administration skills.
• Check out the developerWorks DB2 basics series, a group of articles geared
toward beginning users.
Get products and technologies
• A trial version of DB2 9 is available for free download.
• Download DB2 Express-C, a no-charge version of DB2 Express Edition for the
community that offers the same core data features as DB2 Express Edition and
provides a solid base to build and deploy applications.

About the author

Roman Melnyk
Roman B. Melnyk , Ph.D., is a senior member of the DB2 Information Development
team, specializing in database administration, DB2 utilities, and SQL. During more
than eleven years at IBM, Roman has written and edited numerous DB2 books,
articles, and other related materials. Roman coauthored DB2 Version 8: The Official
Guide, DB2: The Complete Reference, DB2 Fundamentals Certification for Dummies,
and DB2 for Dummies. Roman recently edited Apache Derby -- Off to the Races.

Working with DB2 data

Page 22 of 22 © Copyright IBM Corporation 1994, 2006. All rights reserved.