Flattening a numpy array

Question

I have the following three numpy arrays:

a = np.array([ 1, 2, 3, 4, 2, 3, 4 ])
b = np.array([ [1], [2,3,4], [], [2,3,4] ])
c = np.array([ 1, [2,[3,4]], [], [2,3,4] ])

How can I use a single function f that would work on all three arrays, returning the values in all sublists, in unaltered order and unaltered type?

So the answer should be f(a) == f(b) == f(c) == a.

I found the following trick here (Concatenation of inner lists or ints):

def f(b):
    np.array([a for x in np_array for a in (x if isinstance(x, list) else [x])])

But this does not work for array c.

Indeed some typos: missing return statement, and c = np.array([ [1], [2,[3,4]], [], [2,3,4] ]) — physicalattraction
– physicalattraction, Commented Jul 8, 2014 at 16:12
If you don't care about element-wise comparison, then you could implement something like this generator solution: stackoverflow.com/a/2158532/1634191 — wflynny
– wflynny, Commented Jul 8, 2014 at 16:21

wim · Accepted Answer · 2014-07-08 16:26:25Z

2

For those who don't need to handle nesting, ndarray.flatten does the trick.

For your use case, I'm not aware of a numpy solution, but here is a pure python alternative:

>>> b = np.array([ [1], [2,3,4], [], [2,3,4] ])
>>> from compiler.ast import flatten
>>> np.array(flatten(b))
array([1, 2, 3, 4, 2, 3, 4])

It will only work on python 2.

For python 3 try this:

from collections import Iterable

def flatten(collection):
  for element in collection:
    if isinstance(element, Iterable) and not isinstance(element, str):
      for x in flatten(element):
        yield x
    else:
      yield element

def flatten_array(array):
    return np.array(flatten(array))

answered Jul 8, 2014 at 16:26

wim

368k113 gold badges681 silver badges816 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

physicalattraction Over a year ago

Thanks for the solution with generators. However, it does not seem to work. It does never eneter the for element in collection loop (I placed a print statement in there, which I don't see back). Should I somehow call the flatten generator inan iterative way? How is this done?

ankur vishwakarma · Accepted Answer · 2018-06-15 06:29:18Z

1

In Python3 there is an inbuild function flatten() for that

import numpy as np
arr=np.array([[4,7,-6],[3,0,8],[2,0,-4]])
print(arr)
print(arr.flatten())

answered Jun 15, 2018 at 6:29

ankur vishwakarma

112 bronze badges

Comments

Prajot Kuvalekar · Accepted Answer · 2021-02-03 14:23:37Z

0

ravel() method is more efficient than all

import numpy as np
a = np.array([[1,2,3],[4,5,6]])
print(a.ravel())# flattened array

answered Feb 3, 2021 at 14:23

Prajot Kuvalekar

6,7943 gold badges28 silver badges34 bronze badges

Collectives™ on Stack Overflow

Flattening a numpy array

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related