Numpy 2d Array 从每一行切片不同的元素

Question

Lets say i have a 2D numpy array

a= [[1,2,3]
    [4,5,6]]

i can slice it to select 2 elements form each row:

a[:,0:2]

Output:
[[1,2]
 [4,5]]

but how can i slice different rows individually with different lengths, like 3 elements from first row and two from second

Answer 1

You can create two separate 1D arrays by:

a[0][:3], a[1][:2]

But you cannot make a 2D array with two 1D arrays of different sizes. Note that arrays in numpy are matrices, not 2D lists like Python lists.

On the other hand, you can get a different slice but the same size of each row and make them as a 2D array. For example, assume you have the following 2D array:

a = np.array([[1, 2, 3, 4, 5],[6, 7, 8, 9, 10]])

Now you can select, for example, three elements from the 1st row and three elements from the 2nd row but in a different position:

a[0][:3], a[1][2:5]

Then you can combine them to make a new 2D array:

np.vstack((a[0][:3], a[1][2:5]))

And the result would be:

array([[ 1,  2,  3],
       [ 8,  9, 10]])

And if you want to replace columns and rows you can use transpose or T :

np.vstack((a[0][:3], a[1][2:5])).T

Then you have:

array([[ 1,  8],
       [ 2,  9],
       [ 3, 10]])

• You also can make a new 1D array with the combination of these two arrays:

np.hstack((a[0][:3], a[1][2:5]))

And result is:

array([ 1,  2,  3,  8,  9, 10])