python "in" function doesn't work as expected for arrays

Asked Aug 28 '17 at 19:07

Active Aug 28 '17 at 20:46

Viewed 70 times

Example:

import numpy as np

blah = np.array([['10.12.5.40', 0],
                 ['10.12.5.40', 1],
                 ['10.12.5.40', 2]])

print(['10.12.5.40', 2] in blah)

print(['10.12.5.40', 10] in blah)

The first usage of in works correctly, giving True for an existing element in the array. However, the second usage asks for a non-existing element but True is returned nevertheless.

edited Aug 28 '17 at 19:17

juanpa.arrivillaga

88,713
10
131
172

asked Aug 28 '17 at 19:07

autoencoder

Do you mean _NumPy_ arrays? – Christian Dean Aug 28 '17 at 19:08
1

Used it without array and got correct results – CIsForCookies Aug 28 '17 at 19:09
3

What is `array` here? `numpy.array`? `array.array`? – Martijn Pieters Aug 28 '17 at 19:10
1

I meant numpy.array – autoencoder Aug 28 '17 at 19:12
Because `in` when working with `numpy.ndarray`s is asking `(x == arr).any()`. Fundamentally, you are confused about what consists of an "element" in a `numpy` array (it isn't a "row"), and how elementwise `==` works. – juanpa.arrivillaga Aug 28 '17 at 19:13
Also note, you are creating an array of unicode objects - the `int`s in your second columns get implicitely cast. Honestly, it doesn't look like you should be using `numpy` arrays at all... – juanpa.arrivillaga Aug 28 '17 at 19:14
@juanpa.arrivillaga: numpy arrays are fine, but they should be using a structured dtype here – Eric Aug 29 '17 at 00:02

python "in" function doesn't work as expected for arrays

0 Answers0