Numpy 'where' on string

Question

I would like to use the numpy.where function on a string array. However, I am unsuccessful in doing so. Can someone please help me figure this out?

For example, when I use numpy.where on the following example I get an error:

import numpy as np

A = ['apple', 'orange', 'apple', 'banana']

arr_index = np.where(A == 'apple',1,0)

I get the following:

>>> arr_index
array(0)
>>> print A[arr_index]
>>> apple

However, I would like to know the indices in the string array, A where the string 'apple' matches. In the above string this happens at 0 and 2. However, the np.where only returns 0 and not 2.

So, how do I make numpy.where work on strings? Thanks in advance.

Martin · Accepted Answer · 2021-07-20 15:46:24Z

14

print(a[arr_index])

not array_index!!

a = np.array(['apple', 'orange', 'apple', 'banana'])

arr_index = np.where(a == 'apple')

print(arr_index)

print(a[arr_index])

edited Jul 20, 2021 at 15:46

Martin

1,10114 silver badges17 bronze badges

answered May 20, 2013 at 3:16

Joran Beasley

114k13 gold badges167 silver badges187 bronze badges

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

2 Comments

Rohit Over a year ago

Thanks for the superfast reply! Ah, that is correct. Yes, but what I want are the indices where the string 'apple' matches in the string array. So, I want something like: arr_index = [0,2] # that is 'apple' occurs are positions 0 and 2 in the string array a. However, it only gives me 0.

Rohit Over a year ago

I just corrected the error in my post so that the main question is highlighted and not the stupid error. :-)

Chakresh · Accepted Answer · 2023-01-05 10:35:45Z

1

I believe an easier way is to just do :

A = np.array(['apple', 'orange', 'apple', 'banana'])
arr_index = np.where(A == 'apple')
print(arr_index)

And you get:

(array([0, 2]),)

answered Jan 5, 2023 at 10:35

Chakresh

677 bronze badges

Comments

aperdomo · Accepted Answer · 2020-06-11 10:46:11Z

the thing is that you need to use an array instead of a list to use where properly (also, use True and False instead of 1 and 0 to get a mask to look for the indexes):

A = ['apple', 'orange', 'apple', 'banana']
arr_mask = np.where(np.array(A) == 'apple',True,False)
arr_index = np.arange(0, len(A))[arr_mask]

This way, you'll get arr_index as: np.array([0,2])

Notice that to use the mask arr_mask or the indexes arr_index to look for the values in A, A needs to be an array:

In [55]: A = ['apple', 'orange', 'apple', 'banana'] 
    ...: arr_mask = np.where(np.array(A) == 'apple',True,False) 
    ...: arr_index = np.arange(0, len(A))[arr_mask]                                                             

In [56]: A[arr_mask]                                                                                            
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-56-f8b153319425> in <module>
----> 1 A[arr_mask]

TypeError: only integer scalar arrays can be converted to a scalar index

In [57]: A[arr_index]                                                                                           
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-57-91c260fe71ab> in <module>
----> 1 A[arr_index]

TypeError: only integer scalar arrays can be converted to a scalar index

In [58]: B = np.array(A)                                                                                        

In [59]: B[arr_mask]                                                                                            
Out[59]: array(['apple', 'apple'], dtype='<U6')

In [60]: B[arr_index]                                                                                           
Out[60]: array(['apple', 'apple'], dtype='<U6')

What you are getting using just the list is that the function np.where() doesn't find anywhere where the condition is satisfied. If you try:

A = ['apple', 'orange', 'apple', 'banana']
arr_index = np.where(A == 'orange',1,0)

You'll get again array(0) as the output.

user1595443 · Accepted Answer · 2020-11-09 12:22:48Z

0

Yet another way:

def GetIndexOfStr(npArray,theStr): 
    #npArray is from type of numpy.ndarray where each item is of type np.str
    return np.where(npArray == theStr)[0][0]

A = np.array(['apple', 'orange', 'apple', 'banana'])
print(A[GetIndexOfStr(A,"apple")]) # ==> this will result in "apple"
print(A[GetIndexOfStr(A,"appleX")]) # ==> this will throw IndexError

answered Nov 9, 2020 at 12:22

user1595443

2273 silver badges12 bronze badges

Collectives™ on Stack Overflow

Numpy 'where' on string

4 Answers 4

2 Comments

Comments

Comments

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

2 Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related