I am reading values from a device through a TCP/IP socket that represent an 
ANSI float value. However, the values come across as 2 elements in a ushort 
array. I need to map the 2 values bit patterns into a floating point number 
to get the actual float values being returned.

In C++ I'd finagle it using pointers since a pointer is analogous to an 
array. I could just create a ushort pointer, give it the address of a float 
variable and assign the values to the pointer using array notation. The float 
variable would then have the proper bit pattern.

I don't know how to do that type of thing in C#. Anyone out there know how?

Thanks,

Mike

Hi,

"flyfisher1952"  wrote in message 
news:9569D47D-57DD-40D7-A391-AA6AD2255A23@microsoft.com...

>I am reading values from a device through a TCP/IP socket that represent an
> ANSI float value. However, the values come across as 2 elements in a 
> ushort
> array. I need to map the 2 values bit patterns into a floating point 
> number
> to get the actual float values being returned.
>
> In C++ I'd finagle it using pointers since a pointer is analogous to an
> array. I could just create a ushort pointer, give it the address of a 
> float
> variable and assign the values to the pointer using array notation. The 
> float
> variable would then have the proper bit pattern.
>
> I don't know how to do that type of thing in C#. Anyone out there know 
> how?


Take a look at BitConverter class

Here is the solution using BitConverter.

byte[] bytes = new byte[4];
byte[] highBytes = BitConverter.GetBytes(u1);
byte[] lowBytes = BitConverter.GetBytes(u2);

// The order is funky (little endian), but that's intel for you.
bytes[0] = lowBytes[0];
bytes[1] = lowBytes[1];
bytes[2] = highBytes[0];
bytes[3] = highBytes[1];

float fValue = BitConverter.ToSingle(bytes, 0);

return fValue;

-- 
Jack of All Trades, Master of None


"Ignacio Machin ( .NET/ C# MVP )" wrote:


> Hi,
> 
> "flyfisher1952"  wrote in message 
> news:9569D47D-57DD-40D7-A391-AA6AD2255A23@microsoft.com...
> >I am reading values from a device through a TCP/IP socket that represent an
> > ANSI float value. However, the values come across as 2 elements in a 
> > ushort
> > array. I need to map the 2 values bit patterns into a floating point 
> > number
> > to get the actual float values being returned.
> >
> > In C++ I'd finagle it using pointers since a pointer is analogous to an
> > array. I could just create a ushort pointer, give it the address of a 
> > float
> > variable and assign the values to the pointer using array notation. The 
> > float
> > variable would then have the proper bit pattern.
> >
> > I don't know how to do that type of thing in C#. Anyone out there know 
> > how?
> 
> Take a look at BitConverter class 
> 
> 
>