Numpy mean cannot get other axis-CodePudding

I have my numpy object as:

[array([1, 1, 1]) array([2, 2, 2])
 array([3, 3, 3]) array([4, 4, 4])
 array([5, 5, 5]) array([6, 6, 6])]

My goal is to get a list/array of averages in all sub arrays, i.e. get the list [1 2 3 4 5 6].

I am getting with both np.mean and np.average: [3.5 3.5 3.5] which is the average of all first, all second, all third elements.

I tried added an argument axis=1, but then I am getting the error:

    avg = a.mean(axis)
  File "---/numpy/core/_methods.py", line 138, in _mean
    rcount = _count_reduce_items(arr, axis)
  File "---/numpy/core/_methods.py", line 57, in _count_reduce_items
    items *= arr.shape[ax]
IndexError: tuple index out of range

I tried casting my object to np.array but this made no difference.

As I commented, this is a minimal reproduce of my problem:

import numpy as np
A = np.zeros((2,2), dtype=object)
A[0][0] = np.append(A[0][0], np.array([1]))
A[0][1] = np.append(A[0][1], np.array([2]))
print(np.array(A[0,:]).mean())          # ok
print(np.array(A[0,:]).mean(axis=0))    # ok
np.array(A[0,:]).mean(axis=1)           # bad

CodePudding user response：

I think you wanted to write a.mean(axis=axis). Complete solution:

Code

from numpy import array

a = array(
    [
        array([1, 1, 1]),
        array([2, 2, 2]),
        array([3, 3, 3]),
        array([4, 4, 4]),
        array([5, 5, 5]),
        array([6, 6, 6]),
    ]
)
axis = 1

print(a.mean(axis=axis))

Output

[1. 2. 3. 4. 5. 6.]

CodePudding user response：

>>> obj = [np.array([1, 1, 1]), np.array([2, 2, 2]),  np.array([3, 3, 3]), np.array([4, 4, 4]),  np.array([5, 5, 5]), np.array([6, 6, 6])]
>>> np.array(obj).mean(axis=1).tolist()
[1.0, 2.0, 3.0, 4.0, 5.0, 6.0]

To get it as integers

>>> np.array(obj).mean(axis=1, dtype=np.int32).tolist()
[1, 2, 3, 4, 5, 6]

In your case,

import numpy as np
A = np.zeros((2,2), dtype=object)
A[0] = np.array([1])
A[1] = np.array([2])
A.mean(axis=1).tolist()

output: [1.0, 2.0]