I\'m sitting down to write a massive switch() statement to turn SQL datatypes into CLR datatypes in order to generate classes from MSSQL stored procedures. I\'m using this c
/****** Object: Table [dbo].[DbVsCSharpTypes] Script Date: 03/20/2010 03:07:56 ******/
IF EXISTS (SELECT * FROM sys.objects WHERE object_id = OBJECT_ID(N'[dbo].[DbVsCSharpTypes]')
AND type in (N'U'))
DROP TABLE [dbo].[DbVsCSharpTypes]
GO
/****** Object: Table [dbo].[DbVsCSharpTypes] Script Date: 03/20/2010 03:07:56 ******/
SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE TABLE [dbo].[DbVsCSharpTypes](
[DbVsCSharpTypesId] [int] IDENTITY(1,1) NOT NULL,
[Sql2008DataType] [varchar](200) NULL,
[CSharpDataType] [varchar](200) NULL,
[CLRDataType] [varchar](200) NULL,
[CLRDataTypeSqlServer] [varchar](2000) NULL,
CONSTRAINT [PK_DbVsCSharpTypes] PRIMARY KEY CLUSTERED
(
[DbVsCSharpTypesId] ASC
)WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
) ON [PRIMARY]
GO
SET NOCOUNT ON;
SET XACT_ABORT ON;
GO
SET IDENTITY_INSERT [dbo].[DbVsCSharpTypes] ON;
BEGIN TRANSACTION;
INSERT INTO [dbo].[DbVsCSharpTypes]([DbVsCSharpTypesId], [Sql2008DataType], [CSharpDataType], [CLRDataType], [CLRDataTypeSqlServer])
SELECT 1, N'bigint', N'long', N'Int64, Nullable<Int64>', N'SqlInt64' UNION ALL
SELECT 2, N'binary', N'byte[]', N'Byte[]', N'SqlBytes, SqlBinary' UNION ALL
SELECT 3, N'bit', N'bool', N'Boolean, Nullable<Boolean>', N'SqlBoolean' UNION ALL
SELECT 4, N'char', N'char', NULL, NULL UNION ALL
SELECT 5, N'cursor', NULL, NULL, NULL UNION ALL
SELECT 6, N'date', N'DateTime', N'DateTime, Nullable<DateTime>', N'SqlDateTime' UNION ALL
SELECT 7, N'datetime', N'DateTime', N'DateTime, Nullable<DateTime>', N'SqlDateTime' UNION ALL
SELECT 8, N'datetime2', N'DateTime', N'DateTime, Nullable<DateTime>', N'SqlDateTime' UNION ALL
SELECT 9, N'DATETIMEOFFSET', N'DateTimeOffset', N'DateTimeOffset', N'DateTimeOffset, Nullable<DateTimeOffset>' UNION ALL
SELECT 10, N'decimal', N'decimal', N'Decimal, Nullable<Decimal>', N'SqlDecimal' UNION ALL
SELECT 11, N'float', N'double', N'Double, Nullable<Double>', N'SqlDouble' UNION ALL
SELECT 12, N'geography', NULL, NULL, N'SqlGeography is defined in Microsoft.SqlServer.Types.dll, which is installed with SQL Server and can be downloaded from the SQL Server 2008 feature pack.' UNION ALL
SELECT 13, N'geometry', NULL, NULL, N'SqlGeometry is defined in Microsoft.SqlServer.Types.dll, which is installed with SQL Server and can be downloaded from the SQL Server 2008 feature pack.' UNION ALL
SELECT 14, N'hierarchyid', NULL, NULL, N'SqlHierarchyId is defined in Microsoft.SqlServer.Types.dll, which is installed with SQL Server and can be downloaded from the SQL Server 2008 feature pack.' UNION ALL
SELECT 15, N'image', NULL, NULL, NULL UNION ALL
SELECT 16, N'int', N'int', N'Int32, Nullable<Int32>', N'SqlInt32' UNION ALL
SELECT 17, N'money', N'decimal', N'Decimal, Nullable<Decimal>', N'SqlMoney' UNION ALL
SELECT 18, N'nchar', N'string', N'String, Char[]', N'SqlChars, SqlString' UNION ALL
SELECT 19, N'ntext', NULL, NULL, NULL UNION ALL
SELECT 20, N'numeric', N'decimal', N'Decimal, Nullable<Decimal>', N'SqlDecimal' UNION ALL
SELECT 21, N'nvarchar', N'string', N'String, Char[]', N'SqlChars, SqlStrinG SQLChars is a better match for data transfer and access, and SQLString is a better match for performing String operations.' UNION ALL
SELECT 22, N'nvarchar(1), nchar(1)', N'string', N'Char, String, Char[], Nullable<char>', N'SqlChars, SqlString' UNION ALL
SELECT 23, N'real', N'single', N'Single, Nullable<Single>', N'SqlSingle' UNION ALL
SELECT 24, N'rowversion', N'byte[]', N'Byte[]', NULL UNION ALL
SELECT 25, N'smallint', N'smallint', N'Int16, Nullable<Int16>', N'SqlInt16' UNION ALL
SELECT 26, N'smallmoney', N'decimal', N'Decimal, Nullable<Decimal>', N'SqlMoney' UNION ALL
SELECT 27, N'sql_variant', N'object', N'Object', NULL UNION ALL
SELECT 28, N'table', NULL, NULL, NULL UNION ALL
SELECT 29, N'text', N'string', NULL, NULL UNION ALL
SELECT 30, N'time', N'TimeSpan', N'TimeSpan, Nullable<TimeSpan>', N'TimeSpan' UNION ALL
SELECT 31, N'timestamp', NULL, NULL, NULL UNION ALL
SELECT 32, N'tinyint', N'byte', N'Byte, Nullable<Byte>', N'SqlByte' UNION ALL
SELECT 33, N'uniqueidentifier', N'Guid', N'Guid, Nullable<Guid>', N'SqlGuidUser-defined type(UDT)The same class that is bound to the user-defined type in the same assembly or a dependent assembly.' UNION ALL
SELECT 34, N'varbinary ', N'byte[]', N'Byte[]', N'SqlBytes, SqlBinary' UNION ALL
SELECT 35, N'varbinary(1), binary(1)', N'byte', N'byte, Byte[], Nullable<byte>', N'SqlBytes, SqlBinary' UNION ALL
SELECT 36, N'varchar', NULL, NULL, NULL UNION ALL
SELECT 37, N'xml', NULL, NULL, N'SqlXml'
COMMIT;
RAISERROR (N'[dbo].[DbVsCSharpTypes]: Insert Batch: 1.....Done!', 10, 1) WITH NOWAIT;
GO
SET IDENTITY_INSERT [dbo].[DbVsCSharpTypes] OFF;
This doesn't directly answer the question as asked, but it does answer a common related one. Once you have an IDataReader
you can call IDataRecord.GetFieldType(int) to "[get] the Type
information corresponding to the type of Object
that would be returned from GetValue
."
Normally I just use the Value property to convert a SqlType to a native .NET type. This does the job most of the time. If I have a corner case, I'll just write a quick helper function.
int i = dataReader.GetSqlInt32(0).Value;
Here's a revision that accepts nullable.
public static Type GetClrType(SqlDbType sqlType, bool isNullable)
{
switch (sqlType)
{
case SqlDbType.BigInt:
return isNullable ? typeof(long?) : typeof(long);
case SqlDbType.Binary:
case SqlDbType.Image:
case SqlDbType.Timestamp:
case SqlDbType.VarBinary:
return typeof(byte[]);
case SqlDbType.Bit:
return isNullable ? typeof(bool?) : typeof(bool);
case SqlDbType.Char:
case SqlDbType.NChar:
case SqlDbType.NText:
case SqlDbType.NVarChar:
case SqlDbType.Text:
case SqlDbType.VarChar:
case SqlDbType.Xml:
return typeof(string);
case SqlDbType.DateTime:
case SqlDbType.SmallDateTime:
case SqlDbType.Date:
case SqlDbType.Time:
case SqlDbType.DateTime2:
return isNullable ? typeof(DateTime?) : typeof(DateTime);
case SqlDbType.Decimal:
case SqlDbType.Money:
case SqlDbType.SmallMoney:
return isNullable ? typeof(decimal?) : typeof(decimal);
case SqlDbType.Float:
return isNullable ? typeof(double?) : typeof(double);
case SqlDbType.Int:
return isNullable ? typeof(int?) : typeof(int);
case SqlDbType.Real:
return isNullable ? typeof(float?) : typeof(float);
case SqlDbType.UniqueIdentifier:
return isNullable ? typeof(Guid?) : typeof(Guid);
case SqlDbType.SmallInt:
return isNullable ? typeof(short?) : typeof(short);
case SqlDbType.TinyInt:
return isNullable ? typeof(byte?) : typeof(byte);
case SqlDbType.Variant:
case SqlDbType.Udt:
return typeof(object);
case SqlDbType.Structured:
return typeof(DataTable);
case SqlDbType.DateTimeOffset:
return isNullable ? typeof(DateTimeOffset?) : typeof(DateTimeOffset);
default:
throw new ArgumentOutOfRangeException("sqlType");
}
}
internal Type type(SqlDbType sqltype)
{
Type resulttype = null;
Dictionary<SqlDbType, Type> Types = new Dictionary<SqlDbType, Type>();
Types.Add(SqlDbType.BigInt, typeof(Int64));
Types.Add(SqlDbType.Binary, typeof(Byte[]));
Types.Add(SqlDbType.Bit, typeof(Boolean));
Types.Add(SqlDbType.Char, typeof(String));
Types.Add(SqlDbType.Date, typeof(DateTime));
Types.Add(SqlDbType.DateTime, typeof(DateTime));
Types.Add(SqlDbType.DateTime2, typeof(DateTime));
Types.Add(SqlDbType.DateTimeOffset, typeof(DateTimeOffset));
Types.Add(SqlDbType.Decimal, typeof(Decimal));
Types.Add(SqlDbType.Float, typeof(Double));
Types.Add(SqlDbType.Image, typeof(Byte[]));
Types.Add(SqlDbType.Int, typeof(Int32));
Types.Add(SqlDbType.Money, typeof(Decimal));
Types.Add(SqlDbType.NChar, typeof(String));
Types.Add(SqlDbType.NText, typeof(String));
Types.Add(SqlDbType.NVarChar, typeof(String));
Types.Add(SqlDbType.Real, typeof(Single));
Types.Add(SqlDbType.SmallDateTime, typeof(DateTime));
Types.Add(SqlDbType.SmallInt, typeof(Int16));
Types.Add(SqlDbType.SmallMoney, typeof(Decimal));
Types.Add(SqlDbType.Text, typeof(String));
Types.Add(SqlDbType.Time, typeof(TimeSpan));
Types.Add(SqlDbType.Timestamp, typeof(Byte[]));
Types.Add(SqlDbType.TinyInt, typeof(Byte));
Types.Add(SqlDbType.UniqueIdentifier, typeof(Guid));
Types.Add(SqlDbType.VarBinary, typeof(Byte[]));
Types.Add(SqlDbType.VarChar, typeof(String));
Types.TryGetValue(sqltype, out resulttype);
return resulttype;
}
internal SqlDbType type(Type systype)
{
SqlDbType resulttype = SqlDbType.NVarChar;
Dictionary<Type, SqlDbType> Types = new Dictionary<Type, SqlDbType>();
Types.Add(typeof(Boolean), SqlDbType.Bit);
Types.Add(typeof(String), SqlDbType.NVarChar);
Types.Add(typeof(DateTime), SqlDbType.DateTime);
Types.Add(typeof(Int16), SqlDbType.Int);
Types.Add(typeof(Int32), SqlDbType.Int);
Types.Add(typeof(Int64), SqlDbType.Int);
Types.Add(typeof(Decimal), SqlDbType.Float);
Types.Add(typeof(Double), SqlDbType.Float);
Types.TryGetValue(systype, out resulttype);
return resulttype;
}
I understand that you're discussing writing a switch statement, but here's an alternate for Sql Server (similar concepts work for other DBs)
Consider using SysObjects to retrieve the full data types and generate your class:
declare @ProcName varchar(255)
select @ProcName='Table, View, or Proc'
SELECT --DISTINCT
b.name
, c.name Type
, b.xtype
, b.length
, b.isoutparam
FROM
sysObjects a
INNER JOIN sysCOLUMNs b ON a.id=b.id
INNER JOIN systypes c ON b.xtype=c.xtype
WHERE
a.name=@ProcName
order by b.colorder
Now you're just enumerating a DataTable instead of the longer statement.