1

I want to create a Numpy array, which contains two (Python-) lists. When i try to append an element to one of those lists, the code crashes.

import numpy as np
list0 = [0,0,0]
list1 = [1,1,1]
#list1.append(0)
print(type(list0))
print(type(list1))
array0 = np.array((list0, list1))
array0[0].append(42)
print(array0)

The confusing thing is that when i uncomment the 4th line, the code works just fine.

The error message i get:

File "test.py", line 10, in <module>
array0[0].append(3)
AttributeError: 'numpy.ndarray' object has no attribute 'append'

I run on python 3.5.1 and numpy 1.10.4

The_Mad_Duck
  • 21
  • 1
  • 4
  • I just found out that `np.array(((0,0,0),(1,1,1)))` tries to create a 2d array. This does not happen, if the two lists (e.g. `(0,0,0)` and `(1,1,1)`) have different size (e.g. `(0,0)` and `(1,1,1)`). A way to initialize an array with two empty lists is to write `array0 = np.empty(2, dtype=np.object)` `array0[:] = [], []` – The_Mad_Duck Apr 13 '16 at 09:51
  • Yes, `np.array` defaults to making a mulidimensional array. Making an object dtype is the 2nd class backup choice. It may be better to use plain lists, even faster. – hpaulj Apr 13 '16 at 12:53

2 Answers2

0

How about using numpy's stack functions? You can use vstack (vertical stack) and hstack (horizontal stack) to append lists/arrays together. You can then also continue to stack more lists/arrays onto the newly created stack. I give three examples below.

Python (saved in file stackingArrays)

import numpy as np

list0 = [0,0,0]
list1 = [1,1,1]

# stack vertically
array_v=np.vstack((list0,list1))
print array_v

# stack horizontally
array_h=np.hstack((list0,list1))
print array_h

# stack more on to stacked array
array_v2=np.vstack((array_v,list1))
print array_v2

Output

> python -i stackingArrays.py
>>> [[0 0 0]
    [1 1 1]]
>>> [0 0 0 1 1 1]
>>> [[0 0 0]
    [1 1 1]
    [1 1 1]]
James McCormac
  • 1,635
  • 2
  • 12
  • 26
0

the lists has both the same size that's why it become a 2d np array.

therefor you are trying to append a value to one of the rows of the np array (that you can't because it's not a list any more).

I recommend you to use a np array of lists as shown here.

good luck ;)

itai schwed
  • 399
  • 4
  • 5