如何为两列 numpy 数组中的所有坐标添加偏移量？

Question

The python code below creates data for a numpy array that I use to graph a unit box on a graph:

box = np.array([[x, y] for x in np.arange(0.0, 1.01, 0.01) for y in
               np.arange(0.0, 1.01, 0.01)])

I want to transform box -- by adding a number to the x component and a different number to the y component -- into another numpy array so the new box appears elsewhere on the graph.

I am having some trouble figuring out if I can slice a numpy array to do the addition I need or what the correct loop syntax would be.

My question is: how do I add, say 1 to each x element and 3 to each y element?

So, if some element in the initial numpy array was [0.8, 0.5] , that particular element would then be (in the new array): [1.8, 3.5] . All other elements would also have their x and y values updated the same way.

Answer 1

You can do something like this (just to explain how it works for individual columns):

# data is the array containing x,y values
 
data[:,:1] += 1 # for the first column
data[:,1:] += 3 # for the second column

print(data)

Answer 2

You can use broadcasting. Right now you have an (n, 2) array. You can add a two element array to it directly.

offset = [1., 3.]
box2 = box + offset

This works because dimensions align on the right for broadcasting (and the list offset automatically gets converted to an array). (n, 2) broadcasts just fine with (2,) .

To do the operation in-place (using same memory instead of creating a new output array):

box += offset

While you are at it, you may want to take a look at np.meshgrid and this question for examples of how to create the box much more efficiently than python list comprehensions.