排序非整数2D numpy数组

Question

how would you sort a 2d array with x and y coordinates that are non-integers and are approximate? So, for example such an array as:

[
[0.005, 0.02]
[-0.1, 1.001]
[0.99, 0.004]
[1.1, 0.995]
]

Keep in mind that the [0.005, 0.02] corresponding to the x,y coordinate of [0,0] does not necessarily have the lowest x coordinate or the lowest y coordinate. I have seen how to do it for integers, but I am not sure for this case.

Answer 1

As Sven pointed out, you need a comparison. If you are sorting by distance, then you need to compute at least the square. You can use this:

x = np.array([[1,2],[0.1,0.2],[-1,0.5], [2,2], [0,0]])
x[np.multiply(x,x).sum(axis=1).argsort()]

If you want to sort by x or y, you can use argsort on a slice:

x[x[:,0].argsort()] # sort by x
x[x[:,1].argsort()] # sort by y

Answer 2

You have an (N,2) array of floats. I made a dummy array similar to yours:

>>import numpy as np
>>A = np.random.random((5,2))*3 - 1 
>>A

array([[-0.09759485,  1.09646624],
       [ 1.24045241,  0.59099876],
       [-0.43080349, -0.33879412],
       [ 0.82403019,  0.16274243],
       [ 1.95623418, -0.64082276]])

From what you said, these values are approximate. Before ordering them we can round them to the nearest integers.

>>A = np.round(A)
>>A
array([[-0.,  1.],
       [ 1.,  1.],
       [-0., -0.],
       [ 1.,  0.],
       [ 2., -1.]])

Now numpy.sort() should give you the array ordered how you wanted:

>>np.sort(A, axis=0)
>>A
array([[-0., -1.],
       [-0., -0.],
       [ 1.,  0.],
       [ 1.,  1.],
       [ 2.,  1.]])