Find index of numpy array in list

Question

Can someone explain why the following occurs? My use case is that I have a python list whose elements are all numpy ndarray objects and I need to search through the list to find the index of a particular ndarray obj.

Simplest Example:

>>> import numpy as np
>>> a,b = np.arange(0,5), np.arange(1,6)
>>> a
array([0, 1, 2, 3, 4])
>>> b
array([1, 2, 3, 4, 5])
>>> l = list()
>>> l.append(a)
>>> l.append(b)
>>> l.index(a)
0
>>> l.index(b)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Why can l find the index of a, but not b?

Generally this error arises when you try to perform a Python boolean operation ('in','if') on a numpy boolean array (or list). e.g. if [True,False...]:. Here I suspect one index matches on ids, the other tries an == test. Seems to depend on how the list is constructed. — hpaulj
– hpaulj, Commented May 19, 2015 at 17:18
List index operations involving numpy arrays are probably not a good idea. — hpaulj
– hpaulj, Commented May 19, 2015 at 17:20
I think @hpaulj, got it right. Python probably tries a is comparison first, if that fails python compares with ==. The latter does not work on numpy arrays and gives you the error. Your code probably only works because the very first is comparison is already true and you don't ever invoke a == comparison in this case. — cel
– cel, Commented May 19, 2015 at 17:23
One caveat: if a and b are different lengths, then no error occurs because a == b returns False instead of a bool array. — waldol1
– waldol1, Commented May 19, 2015 at 17:35

Community · Accepted Answer · 2017-05-23 12:06:09Z

9

Applying the idea in https://stackoverflow.com/a/17703076/901925 (see the Related sidebare)

[np.array_equal(b,x) for x in l].index(True)

should be more reliable. It ensures a correct array to array comparison.

Or [id(b)==id(x) for x in l].index(True) if you want to ensure it compares ids.

edited May 23, 2017 at 12:06

CommunityBot

11 silver badge

answered May 19, 2015 at 17:38

hpaulj

233k14 gold badges260 silver badges392 bronze badges

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

Comments

farhawa · Accepted Answer · 2015-05-19 17:38:35Z

0

The idea is to convert numpy arrays to lists and transform the problem to finding a list in an other list:

def find_array(list_of_numpy_array,taregt_numpy_array):
   out = [x.tolist() for x in list_of_numpy_array].index(taregt_numpy_array.tolist())
   return out

answered May 19, 2015 at 17:38

farhawa

10.5k16 gold badges57 silver badges93 bronze badges

Collectives™ on Stack Overflow

Find index of numpy array in list

2 Answers 2

Comments

Comments

Linked

Hot Network Questions