Calculate a ID for a cell based on x and y

I'm calculating a ID for a grid cell like this:

id = x * 65536 + y;

Trying to calculate x and y for a ID I do this:

x = id / 65536;
y开发者_如何学Go = id - (x * 65536);

This works, as long as Y is bigger than 0. If Y is negative, I have to calculate it like this:

x = (int)Math.Ceiling((double)id / 65536.0);
y = id - (x * 65536);

How do I know if I have to round up or down before knowing what y is? Is there a better way to generate one ID for a cell based on x and y? I feel like I'm missing something obvious, but after hours of tracking down strange errors to rounding problems, my brain is not working.

If you know that 30,000 > y > -30,000, you could do this:

x = (id + 30000) / 65536
y = id - (x * 65536)

If 65536 > y > -65536 then:

id = x * 131072 + y

x = (id + 65536) / 131072
y = id - (x * 131702)

You said in the comments that y can take a value from -65336 to 655536. In that case you should use:

id = x * 131072 + y; // 131072 = 2^17

For otherwise let, id be 100000, then this could be translated to two different coordinates:

100000 == 1 * 65536 + 34464  // x = 1, y = 34464
100000 == 2 * 65536 - 31072  // x = 2, y = -31072

So if the range of y is -65536 up to and including 65535, then you should use

id = x * 131072 + y;

and to convert back, you should use the solution that grc proposed:

x = (id + 65536) / 131072;
y = id - (x * 131072);

If C# has modulo (presumably the % operator), use that. (I'm assuming x and y are integers.)

Are you looking for this?

    UInt16 x = 123;
    UInt16 y = 456;

    UInt32 Point;

    // encode y high, x low
    Point = (UInt32)(x | (y << 0x10));

    // decode
    x = (UInt16)(Point & 0x0000FFFF);
    y = (UInt16)((Point & 0xFFFF0000) >> 0x10);