将double [] []类型转换为float

Question

I have a function signature like:

static public double[][] G2Sweep(int row1, int row2, double[][] mtx)
{
    mtx = (float) mtx;
}

But I need to covert the double[][] matrix into float values. How is this done as the code can't explicitly convert it?

Answer 1

public static float[][] Convert(double[][] mtx)
{
    var floatMtx = new float[mtx.Length][];
    for (int i = 0; i < mtx.Length; i++)
    {
        floatMtx[i] = new float[mtx[i].Length];
        for (int j = 0; j < mtx[i].Length; j++)
            floatMtx[i][j] = (float)mtx[i][j];
    }
    return floatMtx;
}

Or:

public static float[][] Convert(double[][] mtx)
{
    return mtx.Select(i => i.Select(j => (float)j).ToArray()).ToArray();
}

Answer 2

No, you can't convert double to float in place, especially for arrays. You need to create copy with correct type.

Answer 3

Another way you can do this by is using the Array.ConvertAll method:

Array.ConvertAll<double, float>(doubles, d => (float)d);

Does the same looping and explicit converting, but looks better.

Answer 4

yes all values to float, but i'd prefer not having to do a for loop to lop through it all

Well, sorry, but you can't. Casting an array to a type like float (or int , or String , or MonkeyPoo , or whatever) makes no sense. If you need to look at every value in an array then you need a loop. There is no getting around that fact.

Lamdas and whatnot all boil down to a loop as well. You'll just need to bite the bullet and either A) Convert in a loop (or use something like .Cast<float> ), or use the correct type to begin with.

Answer 5

It's true that you cant convert in place and that you have to loop through the values.

However, it's fair* to say nowadays that memory is cheap .

At the point of creating your double[][] variable; why not simply also create a float[][] at the same time, so the conversion is done one, in place.

that way in the rest of your lifecycle, you can just use the write array for the write task.

That said, can you clarify why you need a distinct float and double array ?

*Fair but not necessarily acceptable; if it's a webapp hosted on its own box or virtual image; then it's fine. If this is a standalone app that might have to work on a netbook or in Silverlight somewhere, it is not fair or fine.

Answer 6

You can avoid writing the loop by hand if you use Linq.

float[][] floats = mtx.Select(r=>r.Select(Convert.ToSingle).ToArray()).ToArray();

EDIT: fixed.