Understand when to use user defined functions in sql server tech-republic

Understand when to use user-defined functions in SQL Server | TechRepublic

ZDNet Asia SmartPlanet TechRepublic Log In Join TechRepublic FAQ Go Pro!

Blogs Downloads Newsletters Galleries Q&A Discussions News
Research Library

IT Management Development IT Support Data Center Networks Security

Home / Blogs / The Enterprise Cloud Follow this blog:

The Enterprise Cloud

Understand when to use user-
defined functions in SQL
Server
By Tim Chapman
September 3, 2007, 11:49 PM PDT

In the simplest terms, a user-defined function (UDF) in SQL Server is a programming construct
that accepts parameters, does work that typically makes use of the accepted parameters, and
returns a type of result. This article will cover two types of UDFs: table-valued and scalar-valued.
(I will not be covering aggregate functions.)

Types of UDFs
Table-valued functions
A table-valued UDF is a function that accepts parameters and returns the results in the form of a
table. This type of function is special because it returns a table that you can query the results of
and join with other tables. In SQL Server 2005, field values from other tables may be passed into
the function during a join operation to return a record based result. To accomplish this, you must
use SQL Server 2005’s APPLY operator.

It can be difficult to know when it is appropriate to use a VIEW vs. when it is appropriate to use a
table-valued UDF. VIEWs are a great tool for data abstraction, combining data, and logically using
subsets of data. I like to use table-valued UDFs when I need to use one or more values from
different tables in a join operation where some type of calculation needs to be done and an
aggregation returned.

Scalar-valued functions
A scalar-valued UDF accepts parameters and, ultimately, returns a single, atomic value. There are
seven reasons why these types of functions are different than stored procedures in the database
engine.

You cannot modify data inside of a UDF.
A scalar-valued UDF returns only one value, where a stored procedure can have numerous
OUTPUT parameters.
You can use scalar-valued UDFs as the default value for a column in a table.
Scalar-valued UDFs are an easy way to define constant values to use in your database
environment.
You can pass field values as parameters into UDFs.

http://www.techrepublic.com/blog/datacenter/understand-when-to-use-user-defined-functions-in-sql-server/171[08/29/2012 3:08:18 PM]


You can nest scalar function calls. This means that you can pass a call to a scalar-valued
function to another function or stored procedure.
You can use the results from scalar-valued UDFs as criteria in a WHERE statement. Although
you can do it, this is typically not a good idea. (Later in the article, I’ll explain why I try to avoid
this common practice.)

There are two types of scalar-valued UDFs: deterministic and non-deterministic. Recognizing the
determinism of the functions that are created is important. An example of the importance is the
creation of indexed views. One of the many restrictions of creating an index on a view is that the
view definition cannot use a non-deterministic function.

Deterministic
A deterministic UDF always returns the same result with the same set of input parameters. Some
examples of deterministic functions are the system functions MONTH(), YEAR(), and ISNULL().

Non-deterministic
A non-deterministic UDF is can potentially return a different value each time it is called with the
same set of input parameters. Some examples of non-deterministic functions are the system
functions GETDATE(), NEWID(), and @@CONNECTIONS.

Two examples of UDFs
Before presenting the examples, I will set up my SalesHistory table and load data into it:

IF OBJECT_ID('SalesHistory')>0
DROP TABLE SalesHistory;

CREATE TABLE [dbo].[SalesHistory]
(
[SaleID] [int] IDENTITY(1,1) NOT NULL PRIMARY KEY,
[Product] [varchar](10) NULL,
[SaleDate] [datetime] NULL,
[SalePrice] [money] NULL
)

DECLARE @i SMALLINT
SET @i = 1

WHILE (@i <=1000)
BEGIN

INSERT INTO SalesHistory(Product, SaleDate, SalePrice)
VALUES ('Computer', DATEADD(mm, @i, '3/11/1919'), DATEPART(ms, GETDATE()) + (@i +
57))

VALUES('BigScreen', DATEADD(mm, @i, '3/11/1927'), DATEPART(ms, GETDATE()) + (@i +
13))

VALUES('PoolTable', DATEADD(mm, @i, '3/11/1908'), DATEPART(ms, GETDATE()) + (@i +
29))

SET @i = @i + 1
END

GO

The first UDF I will look at is the scalar-valued UDF. The script below defines a function named
dbo.udf_GetProductSales that accepts three parameters and returns a MONEY value. The
function uses the three input parameters as criteria in calculating the total sales from the
SalesHistory table.

CREATE FUNCTION dbo.udf_GetProductSales
(
@Product VARCHAR(10),
@BeginDate DATETIME,
@EndDate DATETIME
)
RETURNS MONEY
AS
BEGIN
DECLARE @Sales MONEY



SELECT @Sales = SUM(SalePrice)
FROM SalesHistory
WHERE
Product = @Product AND
SaleDate BETWEEN @BeginDate AND @EndDate

RETURN(@Sales)
END

The script below calls the UDF created in the above script. Note: The schema the function
belongs to must be used in the call. In this case, the function belongs to the dbo schema.

SELECT dbo.udf_GetProductSales('PoolTable', '1/1/1990', '1/1/2000')

I usually discourage using scalar-valued UDFs in a WHERE criteria statement because, for every
record considered in the query, the scalar-valued function will be called. This means that a
function used in the WHERE criteria will cause a scan of the values being searched, which is
going to be slower than if an index is able to be used. (I will provide more details on this concept
in a future article.)

Although the use of a correlated sub-query is sometimes confusing and complicated, the use of
them can help solve some of the more challenging query problems. While using these special
queries is useful, they only return one column of data. You can use the upgraded table-valued
UDFs in SQL Server 2005 to overcome this shortcoming. I’ll show you how to use the APPLY
operator to accept column values from a table and return a table-result of correlated values.

CREATE FUNCTION dbo.udf_GetProductSalesTable
(
@Product VARCHAR(10),
@SaleID INT
)
RETURNS @SalesTable TABLE
(
SalesTotal MONEY,
SalesCount INT
)

BEGIN

INSERT INTO @SalesTable(SalesTotal, SalesCount)
SELECT
SUM(SalePrice), COUNT(SaleID)
FROM
SalesHistory
WHERE
Product = @Product AND
SaleID <= @SaleID

RETURN
END
GO

The above function accepts the particular product for which we were searching, along with the
SaleID from the SalesHistory table. From the function definition, you can see that the function
returns a table named @SalesTable that contains two columns: SalesTotal and SalesCount. The
body of the function inserts aggregate values into the @SalesTable table variable based upon the
input parameters.

The following code uses the APPLY operator to invoke the table-valued function with the values
from the SalesHistory table. (Note: Logically, you may want to use a JOIN operator here, but it is
not necessary. The APPLY operator essentially does the “JOIN” for us by applying the values from
the SalesHistory table to the table-valued function. In a sense, this code works the same way a
correlated sub-query does, except that it can return multiple correlated values.)

SELECT * FROM SalesHistory sh
CROSS APPLY dbo.udf_GetProductSalesTable(sh.Product, sh.SaleID)
ORDER BY sh.SaleID ASC

Additional TechRepublic resources on UDFs
SQL Server 2000’s user-defined functions add flexibility, cut coding time
Building and using table UDFs in SQL Server
Incorporate SQL Server UDFs in your .NET applications



Tim Chapman is a SQL Server database administrator who works for a bank in Louisville, KY, and
has more than 7 years of IT experience. He is also Microsoft certified in SQL Server 2000 and
SQL Server 2005. If you would like to contact Tim, please e-mail him at chapman.tim@gmail.com.
—————————————————————————————–

Get SQL tips in your inbox
TechRepublic’s free SQL Server newsletter, delivered each Tuesday, contains hands-on tips that
will help you become more adept with this powerful relational database management system.
Automatically subscribe today!

Get IT Tips, news, and reviews delivered directly to your inbox by subscribing to TechRepublic’s free
newsletters.

About Tim Chapman
Full Bio Contact

Talking intelligently about Defragment your Windows
virtualization Server 2003 hard drive on a
set schedule

7 Join the conversation! Add Your Opinion
Comments Follow via:

Staff Picks Top Rated Most Recent My Contacts See All Comments

UDF or View 0
pixelwiz 15th Jul 2011 Votes

Can you please give some more details on when it is more appropriate to use a View
vs a table-valued UDF? I have a few table-valued UDFs created that now I feel
maybe should have been views, but I... Read Whole Comment +

View in thread

Scalar UDF in Where Clause 0
unclebiguns@... 13th Sep 2007 Votes

So you are saying that if I use the a Scalar UDF ona variable or parameter in a where
clause, it will not use an index? For example: Select product_id, product_name from
products where... Read Whole Comment +

View in thread

RE: Understand when to use user-defined functions in 0
SQL Server Votes

alaniane@... 7th Sep 2007

Thanks for the article on UDFs. I also like to use UDFs when designing queries. It



makes it easier to break complex queries into smaller chunks for debugging
purposes. Later on, I will replace the... Read Whole Comment +

View in thread

See all comments

Join the TechRepublic Community and join the conversation! Signing-up is
free and quick, Do it now, we want to hear your opinion.

Join Login


Understand when to use user defined functions in sql server tech-republic

More Related Content

What's hot (20)

Similar to Understand when to use user defined functions in sql server tech-republic (20)

More from Kaing Menglieng (20)

Understand when to use user defined functions in sql server tech-republic