I have a numpy array which looks like
a = ['blue' 'red' 'green']
and I want it to become
b = ['blue', 'red', 'green']
I tried
b = a.split(' ')
but it returns an error: 'numpy.ndarray' object has no attribute 'split'
-
2There's no numpy array in the snippet of code you've posted.jme– jme2015-08-15 16:15:19 +00:00Commented Aug 15, 2015 at 16:15
-
1This is just a display issue.hpaulj– hpaulj2015-08-15 17:33:09 +00:00Commented Aug 15, 2015 at 17:33
-
Does this answer your question? string representation of a numpy array with commas separating its elementsvaragrawal– varagrawal2020-06-25 01:09:19 +00:00Commented Jun 25, 2020 at 1:09
Add a comment
|
3 Answers
Simply turn it to a list:
a = numpy.array(['blue', 'red', 'green'])
print a
>> ['blue' 'red' 'green']
b = list(a)
print b
>> ['blue', 'red', 'green']
But why would you have a numpy array with strings?
1 Comment
record arrays, recarrays, structured arrays... are a very common useage of numpy arrays. Numpy doesn't limit its dtype to one common type. There is extensive documentation in the Numpy and SciPy documentation websites. The only requirement is that each column/field is characterized by one dtype
You can simply call tolist:
import numpy as np
a = np.array(['blue', 'red', 'green'])
b = a.tolist()
print(b)
['blue', 'red', 'green']
answered Aug 15, 2015 at 16:29
Padraic Cunningham
181k30 gold badges264 silver badges327 bronze badges
Comments
I had a similar problem with a list without commas and with arbitrary number of spaces. E.g.:
[2650 20 5]
[2670 5]
[1357 963 355]
I solved it this way:
np.array(re.split("\s+", my_list.replace('[','').replace(']','')), dtype=int)
From the console:
>>> import numpy as np
>>> import re
>>> my_list = '[2650 20 5]'
>>> result = np.array(re.split("\s+", my_list.replace('[','').replace(']','')), dtype=int)
>>> result
array([2650, 20, 5])
>>>
