用非数据集中的值替换numpy数组中的所有非唯一值

Question

I am dealing with large (masked) 2D numpy arrays which originate from country wide raster datasets of 10 to 200 meter resolutions. The arrays are very large and can contain several millions of values.

I would like to perform the following operation on these kinds of arrays in the most efficient way possible:

in_array = numpy.array([[1,2,2],[4,4,6]])
out_array = uniqify(in_array)
print(out_array)
>>>
numpy.array([[1,2,3],[4,5,6]])

or some other combination of numbers. It really does not matter the value, what I care about is that there are NO duplicate values throughout the array. Each cell value must be unique, and the magnitude of the cell must not matter.

This is one way of doing it, but I worry that it may seriously break down for large datasets:

def uniqify(array):
    count = 0
    for i in range(array.shape[0]):
        for j in range(array.shape[1]):
            array[i][j]= count
            count = count+1
    return array


array = np.array([[100,2,3],[4,5,5,],[4,8,7]])
uniqified = uniqify(array)
print(uniqified)

I wonder if there are any ready-made, computationally efficient methods to do this without using nested for loops?

Thanks

Answer 1

你可以用一个简单的

  out_array = np.arange(in_array.size).reshape(in_array.shape)

Answer 2

You can modify an array in place by selecting all values via [:] :

A = np.array([[1,2,2],[4,4,6]])

A[:] = np.arange(A.size).reshape(A.shape)

array([[0, 1, 2],
       [3, 4, 5]])