Is there a way to perform a circular bit shift in C#?

Question

I know that the following is true

int i = 17; //binary 10001
int j = i << 1; //decimal 34, binary 100010

But, if you shift too far, t

Answer 1

I have to admit, I just did a search for "C# bit rotate" and found a link to a page with a Java class that would be easily adapted to C#

I also found this in Google Book which is a C++ function with similar behavior

Answer 2

One year ago I've to implement MD4 for my undergraduate thesis. Here it is my implementation of circular bit shift using a UInt32.

private UInt32 RotateLeft(UInt32 x, Byte n)
{
      return UInt32((x << n) | (x >> (32 - n)));
}

Answer 3

The most famous application is the solution to the Josephus problem (as discussed in Concrete Mathematics, see http://oeis.org/A006257). This is basically a puzzle with no obvious applications. In this video, I demonstrated connections between the second order Josephus problem and complete balanced trees. It's still not an application, but moving slightly in the right direction.

Answer 4

If you know the size of type, you could do something like:

uint i = 17;
uint j = i << 1 | i >> 31;

... which would perform a circular shift of a 32 bit value.

As a generalization to circular shift left n bits, on a b bit variable:

/*some unsigned numeric type*/ input = 17;
var result = input  << n | input  >> (b - n);

---

@The comment, it appears that C# does treat the high bit of signed values differently. I found some info on this here. I also changed the example to use a uint.

Answer 5

Sincce .NET Core 3.0 and up there's BitOperations.RotateLeft() and BitOperations.RotateRight() so you can just use something like

BitOperations.RotateRight(12, 3);
BitOperations.RotateLeft(34L, 5);

In previous versions you can use BitRotator.RotateLeft() and BitRotator.RotateRight() in Microsoft.VisualStudio.Utilities

Answer 6

Just as reference on how to do it, these two functions work perfectly for rotating the bits of 1/2word:

static public uint ShiftRight(uint z_value, int z_shift)
{
    return ((z_value >> z_shift) | (z_value << (16 - z_shift))) & 0x0000FFFF;
}

static public uint ShiftLeft(uint z_value, int z_shift)
{
    return ((z_value << z_shift) | (z_value >> (16 - z_shift))) & 0x0000FFFF;
}

It would be easy to extend it for any given size.