CAST and CONVERT (Transact-SQL)

Converts an expression of one data type to another in SQL Server 2008.

Topic link iconTransact-SQL Syntax Conventions

Syntax

Syntax for CAST:
CAST ( expression AS data_type [ (length ) ])

Syntax for CONVERT:
CONVERT ( data_type [ ( length ) ] , expression [ , style ] )

Arguments

  • expression
    Is any valid expression.

  • data_type
    Is the target data type. This includes xml, bigint, and sql_variant. Alias data types cannot be used. For more information about available data types, see Data Types (Transact-SQL).

  • length
    Is an optional integer that specifies the length of the target data type. The default value is 30.

  • style
    Is an integer expression that specifies how the CONVERT function is to translate expression. If style is NULL, NULL is returned. The range is determined by data_type. For more information, see the Remarks section.

Return Types

Returns expression translated to data_type.

Remarks

Date and Time Styles

When expression is a date or time data type, style can be one of the values shown in the following table. Other values are processed as 0. SQL Server supports the date format in Arabic style by using the Kuwaiti algorithm.

Without century (yy) (1)

With century (yyyy)

Standard

Input/Output (3)

-

0 or 100 (1,2)

Default

mon dd yyyy hh:miAM (or PM)

1

101

U.S.

mm/dd/yyyy

2

102

ANSI

yy.mm.dd

3

103

British/French

dd/mm/yyyy

4

104

German

dd.mm.yy

5

105

Italian

dd-mm-yy

6

106(1)

-

dd mon yy

7

107(1)

-

Mon dd, yy

8

108

-

hh:mi:ss

-

9 or 109 (1,2)

Default + milliseconds

mon dd yyyy hh:mi:ss:mmmAM (or PM)

10

110

USA

mm-dd-yy

11

111

JAPAN

yy/mm/dd

12

112

ISO

yymmdd

yyyymmdd

-

13 or 113 (1,2)

Europe default + milliseconds

dd mon yyyy hh:mi:ss:mmm(24h)

14

114

-

hh:mi:ss:mmm(24h)

-

20 or 120 (2)

ODBC canonical

yyyy-mm-dd hh:mi:ss(24h)

-

21 or 121 (2)

ODBC canonical (with milliseconds)

yyyy-mm-dd hh:mi:ss.mmm(24h)

-

126 (4)

ISO8601

yyyy-mm-ddThh:mi:ss.mmm (no spaces)

-

127(6, 7)

ISO8601 with time zone Z.

yyyy-mm-ddThh:mi:ss.mmmZ

(no spaces)

-

130 (1,2)

Hijri (5)

dd mon yyyy hh:mi:ss:mmmAM

-

131 (2)

Hijri (5)

dd/mm/yy hh:mi:ss:mmmAM

1 These style values return nondeterministic results. Includes all (yy) (without century) styles and a subset of (yyyy) (with century) styles.

2 The default values (style0 or 100, 9 or 109, 13 or 113, 20 or 120, and 21 or 121) always return the century (yyyy).

3 Input when you convert to datetime; output when you convert to character data.

4 Designed for XML use. For conversion from datetime or smalldatetime to character data, the output format is as described in the previous table.

5 Hijri is a calendar system with several variations. SQL Server uses the Kuwaiti algorithm.

Important

By default, SQL Server interprets two-digit years based on a cutoff year of 2049. That is, the two-digit year 49 is interpreted as 2049 and the two-digit year 50 is interpreted as 1950. Many client applications, such as those based on Automation objects, use a cutoff year of 2030. SQL Server provides the two digit year cutoff configuration option that changes the cutoff year used by SQL Server and allows for the consistent treatment of dates. We recommend specifying four-digit years.

6 Only supported when casting from character data to datetime or smalldatetime. When character data that represents only date or only time components is cast to the datetime or smalldatetime data types, the unspecified time component is set to 00:00:00.000, and the unspecified date component is set to 1900-01-01.

7The optional time zone indicator, Z, is used to make it easier to map XML datetime values that have time zone information to SQL Server datetime values that have no time zone. Z is the indicator for time zone UTC-0. Other time zones are indicated with HH:MM offset in the + or - direction. For example: 2006-12-12T23:45:12-08:00.

When you convert to character data from smalldatetime, the styles that include seconds or milliseconds show zeros in these positions. You can truncate unwanted date parts when you convert from datetime or smalldatetime values by using an appropriate char or varchar data type length.

When you convert to datetimeoffset from character data with a style that includes a time, a time zone offset is appended to the result.

float and real Styles

When expression is float or real, style can be one of the values shown in the following table. Other values are processed as 0.

Value

Output

0 (default)

A maximum of 6 digits. Use in scientific notation, when appropriate.

1

Always 8 digits. Always use in scientific notation.

2

Always 16 digits. Always use in scientific notation.

126, 128, 129

Included for legacy reasons and might be deprecated in a future release.

money and smallmoney Styles

When expression is money or smallmoney, style can be one of the values shown in the following table. Other values are processed as 0.

Value

Output

0 (default)

No commas every three digits to the left of the decimal point, and two digits to the right of the decimal point; for example, 4235.98.

1

Commas every three digits to the left of the decimal point, and two digits to the right of the decimal point; for example, 3,510.92.

2

No commas every three digits to the left of the decimal point, and four digits to the right of the decimal point; for example, 4235.9819.

126

Equivalent to style 2 when converting to char(n) or varchar(n)

xml Styles

When expression is xml, style can be one of the values shown in the following table. Other values are processed as 0.

Value

Output

0 (default)

Use default parsing behavior that discards insignificant white space and does not allow for an internal DTD subset.

NoteNote
When you convert to the xml data type, SQL Server insignificant white space is handled differently than in XML 1.0. For more information, see Generating XML Instances.

1

Preserve insignificant white space. This style setting sets the default xml:space handling to behave the same as if xml:space="preserve" has been specified instead.

2

Enable limited internal DTD subset processing.

If enabled, the server can use the following information that is provided in an internal DTD subset to perform nonvalidating parse operations.

  • Defaults for attributes are applied.

  • Internal entity references are resolved and expanded.

  • The DTD content model will be checked for syntactical correctness.

The parser will ignore external DTD subsets. It also does not evaluate the XML declaration to see whether the standalone attribute is set yes or no, but instead parses the XML instance as if it is a stand-alone document.

3

Preserve insignificant white space and enable limited internal DTD subset processing.

Binary Styles

When expression is binary(n), varbinary(n), char(n), or varchar(n), style can be one of the values shown in the following table. Style values that are not listed in the table return an error.

Value

Output

0 (default)

Translates ASCII characters to binary bytes or binary bytes to ASCII characters. Each character or byte is converted 1:1.

If the data_type is a binary type, the characters 0x are added to the left of the result.

1, 2

If the data_type is a binary type, the expression must be a character expression. The expression must be composed of an even number of hexadecimal digits (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F, a, b, c, d, e, f). If the style is set to 1 the characters 0x must be the first two characters in the expression. If the expression contains an odd number of characters or if any of the characters are invalid an error is raised.

If the length of the converted expression is greater than the length of the data_type the result will be right truncated.

Fixed length data_types that are larger then the converted result will have zeros added to the right of the result.

If the data_type is a character type, the expression must be a binary expression. Each binary character is converted into two hexadecimal characters. If the length of the converted expression is greater than the data_type length it will be right truncated.

If the data_type is a fix sized character type and the length of the converted result is less than its length of the data_type; spaces are added to the right of the converted expression to maintain an even number of hexadecimal digits.

The characters 0x will be added to the left of the converted result for style 1.

Implicit Conversions

Implicit conversions are those conversions that occur without specifying either the CAST or CONVERT function. Explicit conversions are those conversions that require the CAST or CONVERT function to be specified. The following illustration shows all explicit and implicit data type conversions that are allowed for SQL Server system-supplied data types. These include xml, bigint, and sql_variant. There is no implicit conversion on assignment from the sql_variant data type, but there is implicit conversion to sql_variant.

Data type conversion table

When you convert between datetimeoffset and the character types char, varchar, nchar, and nvarchar the converted time zone offset part should always be double digits for both HH and MM for example, -08:00.

Note

Because Unicode data always uses an even number of bytes, use caution when you convert binary or varbinary to or from Unicode supported data types. For example, the following conversion does not return a hexadecimal value of 41; it returns 4100: SELECT CAST(CAST(0x41 AS nvarchar) AS varbinary).

Large-Value Data Types

Large-value data types exhibit the same implicit and explicit conversion behavior as their smaller counterparts, specifically the varchar, nvarchar and varbinary data types. However, you should consider the following guidelines:

  • Conversion from image to varbinary(max) and vice-versa is an implicit conversion, and so are conversions between text and varchar(max), and ntext and nvarchar(max).

  • Conversion from large-value data types, such as varchar(max), to a smaller counterpart data type, such as varchar, is an implicit conversion, but truncation will occur if the large value is too big for the specified length of the smaller data type.

  • Conversion from varchar, nvarchar, or varbinary to their corresponding large-value data types is performed implicitly.

  • Conversion from the sql_variant data type to the large-value data types is an explicit conversion.

  • Large-value data types cannot be converted to the sql_variant data type.

For information about how to convert Microsoft .NET Framework common language runtime (CLR) user-defined types, see Performing Operations on User-defined Types. For more information about how to convert from the xml data type, see Generating XML Instances.

xml Data Type

When you explicitly or implicitly cast the xml data type to a string or binary data type, the content of the xml data type is serialized based on a set of rules. For information about these rules, see Serialization of XML Data. For information on how to cast from XML to a CLR user-defined type, see Performing Operations on User-defined Types. For information about how to convert from other data types to the xml data type, see Generating XML Instances.

text and image Data Types

Automatic data type conversion is not supported for the text and image data types. You can explicitly convert text data to character data, and image data to binary or varbinary, but the maximum length is 8000 bytes. If you try an incorrect conversion such as trying to convert a character expression that includes letters to an int, SQL Server returns an error message.

Output Collation

When the output of CAST or CONVERT is a character string, and the input is a character string, the output has the same collation and collation label as the input. If the input is not a character string, the output has the default collation of the database, and a collation label of coercible-default. For more information, see Collation Precedence (Transact-SQL).

To assign a different collation to the output, apply the COLLATE clause to the result expression of the CAST or CONVERT function. For example:

SELECT CAST('abc' AS varchar(5)) COLLATE French_CS_AS

Truncating and Rounding Results

When you convert character or binary expressions (char, nchar, nvarchar, varchar, binary, or varbinary) to an expression of a different data type, data can be truncated, only partially displayed, or an error is returned because the result is too short to display. Conversions to char, varchar, nchar, nvarchar, binary, and varbinary are truncated, except for the conversions shown in the following table.

From data type

To data type

Result

int, smallint, or tinyint

char

*

 

varchar

*

 

nchar

E

 

nvarchar

E

money, smallmoney, numeric, decimal, float, or real

char

E

 

varchar

E

 

nchar

E

 

nvarchar

E

* = Result length too short to display. E = Error returned because result length is too short to display.

SQL Server guarantees that only roundtrip conversions, conversions that convert a data type from its original data type and back again, will yield the same values from version to version. The following example shows such a roundtrip conversion:

DECLARE @myval decimal (5, 2)
SET @myval = 193.57
SELECT CAST(CAST(@myval AS varbinary(20)) AS decimal(10,5))
-- Or, using CONVERT
SELECT CONVERT(decimal(10,5), CONVERT(varbinary(20), @myval))

Note

Do not try to construct binary values and then convert them to a data type of the numeric data type category. SQL Server does not guarantee that the result of a decimal or numeric data type conversion to binary will be the same between versions of SQL Server.

The following example shows a resulting expression that is too small to display.

USE AdventureWorks;
GO
SELECT c.FirstName, c.LastName, SUBSTRING(c.Title, 1, 25) AS Title, CAST(e.SickLeaveHours AS char(1)) AS 'Sick Leave'
FROM HumanResources.Employee e JOIN Person.Contact c ON e.EmployeeID = c. ContactID
WHERE NOT EmployeeID >5

Here is the result set.

FirstName LastName       Title                Sick Leave
-------------- -------------  --------------------  -----------
Gustavo        Achong         Mr.                   *
Catherine      Abel           Ms.                   *
Kim            Abercrombie    Ms.                   *
Humberto       Acevedo        Sr.                   *
Pilar          Ackerman       Sra.                  *

(5 row(s) affected)

When you convert data types that differ in decimal places, sometimes the result value is truncated and at other times it is rounded. The following table shows the behavior.

From

To

Behavior

numeric

numeric

Round

numeric

int

Truncate

numeric

money

Round

money

int

Round

money

numeric

Round

float

int

Truncate

float

numeric

Round

float

datetime

Round

datetime

int

Round

For example, the result of the following conversion is 10:

SELECT CAST(10.6496 AS int)

When you convert data types in which the target data type has fewer decimal places than the source data type, the value is rounded. For example, the result of the following conversion is $10.3497:

SELECT CAST(10.3496847 AS money)

SQL Server returns an error message when nonnumeric char, nchar, varchar, or nvarchar data is converted to int, float, numeric, or decimal. SQL Server also returns an error when an empty string (" ") is converted to numeric or decimal.

Certain datetime Conversions Are Nondeterministic in SQL Server 2005 and Later Versions

In SQL Server 2000, string to date and time conversions are marked as deterministic. However, this is not true for the styles listed in the following table. For these styles, the conversions depend on the language settings. SQL Server 2005 and later versions mark these conversions as nondeterministic.

The following table lists the styles for which the string-to-datetime conversion is nondeterministic.

All styles below 1001

106

107

109

113

130

1 With the exception of styles 20 and 21

Examples

A. Using both CAST and CONVERT

Each example retrieves the name of the product for those products that have a 3 in the first digit of their list price and converts their ListPrice to int.

-- Use CAST
USE AdventureWorks;
GO
SELECT SUBSTRING(Name, 1, 30) AS ProductName, ListPrice
FROM Production.Product
WHERE CAST(ListPrice AS int) LIKE '3%';
GO

-- Use CONVERT.
USE AdventureWorks;
GO
SELECT SUBSTRING(Name, 1, 30) AS ProductName, ListPrice
FROM Production.Product
WHERE CONVERT(int, ListPrice) LIKE '3%';
GO

B. Using CAST with arithmetic operators

The following example calculates a single column computation (Computed) by dividing the total year-to-date sales (SalesYTD) by the commission percentage (CommissionPCT). This result is converted to an int data type after being rounded to the nearest whole number.

USE AdventureWorks;
GO
SELECT CAST(ROUND(SalesYTD/CommissionPCT, 0) AS int) AS 'Computed'
FROM Sales.SalesPerson 
WHERE CommissionPCT != 0;
GO

Here is the result set.

Computed

------

379753754

346698349

257144242

176493899

281101272

0

301872549

212623750

298948202

250784119

239246890

101664220

124511336

97688107

(14 row(s) affected)

C. Using CAST to concatenate

The following example concatenates noncharacter, nonbinary expressions by using CAST.

USE AdventureWorks;
GO
SELECT 'The list price is ' + CAST(ListPrice AS varchar(12)) AS ListPrice
FROM Production.Product
WHERE ListPrice BETWEEN 350.00 AND 400.00;
GO

Here is the result set.

ListPrice

------------------

The list price is 357.06

The list price is 364.09

The list price is 364.09

The list price is 364.09

The list price is 364.09

(5 row(s) affected)

D. Using CAST to produce more readable text

The following example uses CAST in the select list to convert the Name column to a char(10) column.

USE AdventureWorks;
GO
SELECT DISTINCT CAST(p.Name AS char(10)) AS Name, s.UnitPrice
FROM Sales.SalesOrderDetail s JOIN Production.Product p on s.ProductID = p.ProductID
WHERE Name LIKE 'Long-Sleeve Logo Jersey, M';
GO

Here is the result set.

Name       UnitPrice

---------- ---------------------

Long-Sleev 31.2437

Long-Sleev 32.4935

Long-Sleev 49.99

(3 row(s) affected)

E. Using CAST with the LIKE clause

The following example converts the money column SalesYTD to an int and then to a char(20) column so that it can be used with the LIKE clause.

USE AdventureWorks;
GO
SELECT p.FirstName, p.LastName, s.SalesYTD, s.SalesPersonID
FROM Person.Contact p JOIN Sales.SalesPerson s ON p.ContactID = s.SalesPersonID
WHERE CAST(CAST(s.SalesYTD AS int) AS char(20)) LIKE '2%';
GO

Here is the result set.

FirstName        LastName            SalesYTD         SalesPersonID

---------------- ------------------- ---------------- -------------

Carol             Elliott             2811012.7151     279

Julie             Estes               219088.8836      288

Janeth            Esteves             2241204.0424     289

(3 row(s) affected)

F. Using CONVERT or CAST with typed XML

The following are several examples that show using CONVERT to convert to typed XML by using the xml data type.

This example converts a string with white space, text and markup into typed XML and removes all insignificant white space (boundary white space between nodes):

CONVERT(XML, '<root><child/></root>')

This example converts a similar string with white space, text and markup into typed XML and preserves insignificant white space (boundary white space between nodes):

CONVERT(XML, '<root>          <child/>         </root>', 1)

This example casts a string with white space, text, and markup into typed XML:

CAST('<Name><FName>Carol</FName><LName>Elliot</LName></Name>'  AS XML)

For more examples, see Generating XML Instances.

G. Using CAST and CONVERT with datetime data

The following example displays the current date and time, uses CAST to change the current date and time to a character data type, and then uses CONVERT display the date and time in the ISO 8901 format.

SELECT 
   GETDATE() AS UnconvertedDateTime,
   CAST(GETDATE() AS nvarchar(30)) AS UsingCast,
   CONVERT(nvarchar(30), GETDATE(), 126) AS UsingConvertTo_ISO8601  ;
GO

Here is the result set.

UnconvertedDateTime     UsingCast                      UsingConvertTo_ISO8601

----------------------- ------------------------------ ------------------------------

2006-04-18 09:58:04.570 Apr 18 2006 9:58AM            2006-04-18T09:58:04.570

(1 row(s) affected)

The following example is approximately the opposite of the previous example. The example displays a date and time as character data, uses CAST to change the character data to the datetime data type, and then uses CONVERT to change the character data to the datetime data type.

SELECT 
   '2006-04-25T15:50:59.997' AS UnconvertedText,
   CAST('2006-04-25T15:50:59.997' AS datetime) AS UsingCast,
   CONVERT(datetime, '2006-04-25T15:50:59.997', 126) AS UsingConvertFrom_ISO8601 ;
GO

Here is the result set.

UnconvertedText         UsingCast               UsingConvertFrom_ISO8601

----------------------- ----------------------- ------------------------

2006-04-25T15:50:59.997 2006-04-25 15:50:59.997 2006-04-25 15:50:59.997

(1 row(s) affected)

H. Using CONVERT with binary and character data

The following examples show the results of converting binary and character data by using different styles.

--Convert the binary value 0x4E616d65 to a character value.
SELECT CONVERT(char(8), 0x4E616d65, 0) AS 'Style 0, binary to character'

Here is the result set.

Style 0, binary to character

----------------------------

Name

(1 row(s) affected)

--The following example shows how Style 1 can force the result
--to be truncated. The truncation is caused by
--including the characters 0x in the result.
SELECT CONVERT(char(8), 0x4E616d65, 1) AS 'Style 1, binary to character'

Here is the result set.

Style 1, binary to character

------------------------------

0x4E616D

(1 row(s) affected)

--The following example shows that Style 2 does not truncate the
--result because the characters 0x are not included in
--the result.
SELECT CONVERT(char(8), 0x4E616d65, 2) AS 'Style 2, binary to character'

Here is the result set.

Style 2, binary to character

------------------------------

4E616D65

(1 row(s) affected)

--Convert the character value 'Name' to a binary value.
SELECT CONVERT(binary(8), 'Name', 0) AS 'Style 0, character to binary'

Here is the result set.

Style 0, character to binary

----------------------------------

0x4E616D6500000000

(1 row(s) affected)

SELECT CONVERT(binary(4), '0x4E616D65', 1) AS 'Style 1, character to binary'

Here is the result set.

Style 1, character to binary

----------------------------------

0x4E616D65

(1 row(s) affected)

SELECT CONVERT(binary(4), '4E616D65', 1) AS 'Style 2, character to binary'

Here is the result set.

Style 2, character to binary

----------------------------------

0x4E616D65

(1 row(s) affected)