将数组的numpy数组转换为2D numpy数组

Question

I have a numpy array of numpy arrays, and I want to convert it to a 2d numpy array. For example

#Init Sub Arrays
a = np.array([1,2])
b = np.array([3,4,5])
c = np.array([2,1])
d = np.array([5,4,3])

#Combine Sub Arrays
e = np.array([[a,b],[c,d]])

#Sample Sub Arrays
f = e[:,0]
#Attempt to convert sub np arrays to 2d np array
g = np.array(f)

expected = np.array([[1,2],[2,1]])
print("Actual 1: ",f)
print("Actual 2: ",g)
print("Expected:", expected)

print("Actual 1: ",np.ravel(f))
print("Actual 2: ",np.ravel(g))
print("Expected: ",np.ravel(expected))

Output:

Actual 1:  [array([1, 2]) array([2, 1])]
Actual 2:  [array([1, 2]) array([2, 1])]
Expected: [[1 2]
 [2 1]]
Actual 1:  [array([1, 2]) array([2, 1])]
Actual 2:  [array([1, 2]) array([2, 1])]
Expected:  [1 2 2 1]

I understand that the array is initialized the way it is because numpy doesn't support arrays of different length on the same dimension, but I want to know how I can convert a valid sample of a "hack" numpy array to a "valid" numpy array

Answer 1

You could just use np.vstack :

out = np.vstack(e[:,0])
print(out)
array([[1, 2],
       [2, 1]])

print(out.ravel())
array([1, 2, 2, 1])