If you have a float in MSSQLServer, to what do you map this in .NET?
Can you convert it to Double or will you lose numbers?
Add a comment
|
6 Answers
SQLServer float and C#/VB double have the same representation. This is the correct mapping. What you don't want to do is map SQL Server float to C#/VB float as that may involve a loss of precision. SQL Server real maps onto C#/VB float.
T-SQL float and real type definitions can be found at MSDN. C# double definition can be found at MSDN as well, as can the float definition.
2 Comments
Stefan Steiger
Wrong, though you can map any SQL-server float to a C# double, you can't map any C# double to a SQL-server float... But it usually works for "normally" large/tiny numbers
user2246674
This isn't strictly true - even for REAL/FLOAT(24). SQL Server has no notion of NaN, or +/-Inf which are supported under standard IEEE754 (C#/VB.NET) rules.
Check out:
- SQL Server Data Types and Their .NET Framework Equivalents (msdn) for full reference on data type mappings between SQL Server and .Net
In your case SQL Server native type float maps to SQL Server CLR SqlDouble and then Double in .Net
Update (May 2016):
Updated version of the msdn document: Mapping CRL Parameter Data
2 Comments
Matthis Kohli
This should be the correct answer. Thanks for the link. There is an updated version by now. msdn.microsoft.com/en-us/library/ms131092%28v=sql.130%29.aspx
kristof
thanks @MatthisKohli - added the new link to the answer now
C# has a float type, but it will convert to double just fine as well.
Comments
From my recollection, most of the ORM tools will map it to a Decimal type, which will not lose precision like a double or float.
Comments
It depends on the size of the SQL float. For plain "float" which is equivalent to float(53), you need to use a C# double. For float(24) or lower a C# float will be enough, or a double would work as well.
Comments
The automatic code generated by Microsoft XSD convert SQL float (default float) to C# double.
So you can suggest that it isn't big mistake.
Until you working with float that have default size.
