CREATE TYPE (Transact-SQL)

**APPLIES TO:** ![yes](media/yes.png)SQL Server (starting with 2008) ![yes](media/yes.png)Azure SQL Database ![no](media/no.png)Azure SQL Data Warehouse ![no](media/no.png)Parallel Data Warehouse

Creates an alias data type or a user-defined type in the current database in SQL Server or SQL Server Azure SQL Database The implementation of an alias data type is based on a SQL Server Azure SQL Database SQL Server native system type. A user-defined type is implemented through a class of an assembly in the SQL Server Azure SQL Database SQL Server Microsoft dnprdnshort] common language runtime (CLR). To bind a user-defined type to its implementation, the CLR assembly that contains the implementation of the type must first be registered in SQL Server Azure SQL Database SQL Server Microsoft SQL Server by using CREATE ASSEMBLY.

The ability to run CLR code is off by default in SQL Server You can create, modify and drop database objects that reference managed code modules, but these references will not execute in SQL Server SQL Server unless the clr enabled Option is enabled by using sp_configure.

[!NOTE]
The integration of .NET Framework CLR into SQL Server is discussed in this topic. CLR integration does not apply to Azure SQL Database

Topic link icon Transact-SQL Syntax Conventions

Syntax

-- User-defined Data Type Syntax    
CREATE TYPE [ schema_name. ] type_name  
{   
    FROM base_type   
    [ ( precision [ , scale ] ) ]  
    [ NULL | NOT NULL ]   
  | EXTERNAL NAME assembly_name [ .class_name ]   
  | AS TABLE ( { <column_definition> | <computed_column_definition> }  
        [ <table_constraint> ] [ ,...n ] )    
} [ ; ]  
  
<column_definition> ::=  
column_name <data_type>  
    [ COLLATE collation_name ]   
    [ NULL | NOT NULL ]  
    [   
        DEFAULT constant_expression ]   
      | [ IDENTITY [ ( seed ,increment ) ]   
    ]  
    [ ROWGUIDCOL ] [ <column_constraint> [ ...n ] ]   
  
<data type> ::=   
[ type_schema_name . ] type_name   
    [ ( precision [ , scale ] | max |   
                [ { CONTENT | DOCUMENT } ] xml_schema_collection ) ]   
  
<column_constraint> ::=   
{     { PRIMARY KEY | UNIQUE }   
        [ CLUSTERED | NONCLUSTERED ]   
        [   
            WITH ( <index_option> [ ,...n ] )   
        ]  
  | CHECK ( logical_expression )   
}   
  
<computed_column_definition> ::=  
  
column_name AS computed_column_expression   
[ PERSISTED [ NOT NULL ] ]  
[   
    { PRIMARY KEY | UNIQUE }  
        [ CLUSTERED | NONCLUSTERED ]  
        [   
            WITH ( <index_option> [ ,...n ] )  
        ]  
    | CHECK ( logical_expression )   
]   
  
<table_constraint> ::=  
{   
    { PRIMARY KEY | UNIQUE }   
        [ CLUSTERED | NONCLUSTERED ]   
    ( column [ ASC | DESC ] [ ,...n ] )   
        [   
    WITH ( <index_option> [ ,...n ] )   
        ]  
    | CHECK ( logical_expression )   
}   
  
<index_option> ::=  
{  
    IGNORE_DUP_KEY = { ON | OFF }  
}

-- User-defined Table Types Syntax  
CREATE TYPE [schema_name. ] type_name  
AS TABLE ( { <column_definition> }  
    |  [ <table_constraint> ] [ ,... n ]    
    | [ <table_index> ] [ ,... n ]    } )
    [ WITH ( <table_option> [ ,... n ] ) ]  
 [ ; ]  
  
<column_definition> ::=  
column_name <data_type>  
    [ COLLATE collation_name ]   [ NULL | NOT NULL ]    [  
      [ IDENTITY [ (1 , 1) ]  
    ]  
    [ <column_constraint> [ ... n ] ]    [ <column_index> ]  
  
<data type> ::=  
 [type_schema_name . ] type_name [ (precision [ , scale ]) ]  
  
<column_constraint> ::=  
{ PRIMARY KEY {   NONCLUSTERED HASH WITH (BUCKET_COUNT = bucket_count) 
                | NONCLUSTERED } }  
  
< table_constraint > ::=  
{ PRIMARY KEY { NONCLUSTERED HASH (column [ ,... n ] ) 
                   WITH (BUCKET_COUNT = bucket_count) 
               | NONCLUSTERED  (column [ ASC | DESC ] [ ,... n ] )  } }  
  
<column_index> ::=  
  INDEX index_name  
{ { [ NONCLUSTERED ] HASH WITH (BUCKET_COUNT = bucket_count) 
     | NONCLUSTERED } }  
  
< table_index > ::=  
  INDEX constraint_name  
{ { [ NONCLUSTERED ] HASH (column [ ,... n ] ) WITH (BUCKET_COUNT = bucket_count) 
 |  [NONCLUSTERED]  (column [ ASC | DESC ] [ ,... n ] )} }  
  
<table_option> ::=  
{  
    [MEMORY_OPTIMIZED = {ON | OFF}]  
}

Arguments

schema_name
Is the name of the schema to which the alias data type or user-defined type belongs.

type_name
Is the name of the alias data type or user-defined type. Type names must comply with the rules for identifiers.

base_type
Is the SQL Server supplied data type on which the alias data type is based. base_type is sysname, with no default, and can be one of the following values:

bigint	binary( n )	bit	char( n )
date	datetime	datetime2	datetimeoffset
decimal	float	image	int
money	nchar( n )	ntext	numeric
nvarchar( n \| max)	real	smalldatetime	smallint
smallmoney	sql_variant	text	time
tinyint	uniqueidentifier	varbinary( n \| max)	varchar( n \| max)

base_type can also be any data type synonym that maps to one of these system data types.

precision
For decimal or numeric, is a non-negative integer that indicates the maximum total number of decimal digits that can be stored, both to the left and to the right of the decimal point. For more information, see decimal and numeric (Transact-SQL).

scale
For decimal or numeric, is a non-negative integer that indicates the maximum number of decimal digits that can be stored to the right of the decimal point, and it must be less than or equal to the precision. For more information, see decimal and numeric (Transact-SQL).

NULL | NOT NULL
Specifies whether the type can hold a null value. If not specified, NULL is the default.

assembly_name
Applies to: SQL Server 2008 through SQL Server 2008 SQL Server 2017

Specifies the SQL Server assembly that references the implementation of the user-defined type in the common language runtime. assembly_name should match an existing assembly in SQL Server SQL Server in the current database.

[!NOTE]
EXTERNAL_NAME is not available in a contained database.

[.** class_name **]
Applies to: SQL Server 2008 through SQL Server 2008 SQL Server 2017

Specifies the class within the assembly that implements the user-defined type. class_name must be a valid identifier and must exist as a class in the assembly with assembly visibility. class_name is case-sensitive, regardless of the database collation, and must exactly match the class name in the corresponding assembly. The class name can be a namespace-qualified name enclosed in square brackets ([ ]) if the programming language that is used to write the class uses the concept of namespaces, such as C#. If class_name is not specified, SQL Server assumes it is the same as type_name.

<column_definition>
Defines the columns for a user-defined table type.

<data type>
Defines the data type in a column for a user-defined table type. For more information about data types, see Data Types (Transact-SQL). For more information about tables, see CREATE TABLE (Transact-SQL).

<column_constraint>
Defines the column constraints for a user-defined table type. Supported constraints include PRIMARY KEY, UNIQUE, and CHECK. For more information about tables, see CREATE TABLE (Transact-SQL).

<computed_column_definition>
Defines a computed column expression as a column in a user-defined table type. For more information about tables, see CREATE TABLE (Transact-SQL).

<table_constraint>
Defines a table constraint on a user-defined table type. Supported constraints include PRIMARY KEY, UNIQUE, and CHECK.

<index_option>
Specifies the error response to duplicate key values in a multiple-row insert operation on a unique clustered or unique nonclustered index. For more information about index options, see CREATE INDEX (Transact-SQL).

INDEX
You must specify column and table indexes as part of the CREATE TABLE statement. CREATE INDEX and DROP INDEX are not supported for memory-optimized tables.

MEMORY_OPTIMIZED
Applies to: SQL Server 2014 (12.x) through SQL Server 2014 (12.x) SQL Server 2017 and SQL Server 2014 (12.x) SQL Server 2017 Azure SQL Database

Indicates whether the table type is memory optimized. This option is off by default; the table (type) is not a memory optimized table (type). Memory optimized table types are memory-optimized user tables, the schema of which is persisted on disk similar to other user tables.

BUCKET_COUNT
Applies to: SQL Server 2014 (12.x) through SQL Server 2014 (12.x) SQL Server 2017 and SQL Server 2014 (12.x) SQL Server 2017 Azure SQL Database

Indicates the number of buckets that should be created in the hash index. The maximum value for BUCKET_COUNT in hash indexes is 1,073,741,824. For more information about bucket counts, see Indexes for Memory-Optimized Tables. bucket_count is a required argument.

HASH
Applies to: SQL Server 2014 (12.x) through SQL Server 2014 (12.x) SQL Server 2017 and SQL Server 2014 (12.x) SQL Server 2017 Azure SQL Database

Indicates that a HASH index is created. Hash indexes are supported only on memory optimized tables.

Remarks

The class of the assembly that is referenced in assembly_name, together with its methods, should satisfy all the requirements for implementing a user-defined type in SQL Server For more information about these requirements, see CLR User-Defined Types.

Additional considerations include the following:

The class can have overloaded methods, but these methods can be called only from within managed code, not from Transact\-SQL
Any static members must be declared as const or readonly if assembly_name is SAFE or EXTERNAL_ACCESS.

Within a database, there can be only one user-defined type registered against any specified type that has been uploaded in SQL Server from the CLR. If a user-defined type is created on a CLR type for which a user-defined type already exists in the database, CREATE TYPE fails with an error. This restriction is required to avoid ambiguity during SQL Type resolution if a CLR type can be mapped to more than one user-defined type.

If any mutator method in the type does not return void, the CREATE TYPE statement does not execute.

To modify a user-defined type, you must drop the type by using a DROP TYPE statement and then re-create it.

Unlike user-defined types that are created by using sp_addtype, the public database role is not automatically granted REFERENCES permission on types that are created by using CREATE TYPE. This permission must be granted separately.

In user-defined table types, structured user-defined types that are used in column_name <data type> are part of the database schema scope in which the table type is defined. To access structured user-defined types in a different scope within the database, use two-part names.

In user-defined table types, the primary key on computed columns must be PERSISTED and NOT NULL.

Memory-Optimized Table Types

Beginning in SQL Server 2014 (12.x) processing data in a table type can be done in primary memory, and not on disk. For more information, see In-Memory OLTP (In-Memory Optimization). For code samples showing how to create memory-optimized table types, see Creating a Memory-Optimized Table and a Natively Compiled Stored Procedure.

Permissions

Requires CREATE TYPE permission in the current database and ALTER permission on schema_name. If schema_name is not specified, the default name resolution rules for determining the schema for the current user apply. If assembly_name is specified, a user must either own the assembly or have REFERENCES permission on it.

If any columns in the CREATE TABLE statement are defined to be of a user-defined type, REFERENCES permission on the user-defined type is required.

[!NOTE] A user creating a table with a column that uses a user-defined type needs the REFERENCES permission on the user-defined type. If this table must be created in TempDB, then either the REFERENCES permission needs to be granted explicitly each time before the table is created, or this data type and REFERENCES permissions need to be added to the Model database. If this is done, then this data type and permissions will be available in TempDB permanently. Otherwise, the user-defined data type and permissions will disappear when SQL Server is restarted. For more information, see CREATE TABLE

Examples

A. Creating an alias type based on the varchar data type

The following example creates an alias type based on the system-supplied varchar data type.

CREATE TYPE SSN  
FROM varchar(11) NOT NULL ;

B. Creating a user-defined type

The following example creates a type Utf8String that references class utf8string in the assembly utf8string. Before creating the type, assembly utf8string is registered in the local database. Replace the binary portion of the CREATE ASSEMBLY statement with a valid description.

Applies to: SQL Server 2008 through SQL Server 2008 SQL Server 2017

CREATE ASSEMBLY utf8string  
AUTHORIZATION [dbi]   
FROM 0x4D... ;  
GO  
CREATE TYPE Utf8String   
EXTERNAL NAME utf8string.[Microsoft.Samples.SqlServer.utf8string] ;  
GO

C. Creating a user-defined table type

The following example creates a user-defined table type that has two columns. For more information about how to create and use table-valued parameters, see Use Table-Valued Parameters (Database Engine).

CREATE TYPE LocationTableType AS TABLE   
    ( LocationName VARCHAR(50)  
    , CostRate INT );  
GO