I have a numpy array like this:
a = [0,88,26,3,48,85,65,16,97,83,91]
How can I get the values at certain index positions in ONE step? For example:
ind_pos = [1,5,7]
The result should be:
[88,85,16]
-
2a is not a numpy array in the question but a simple listbugmenot123– bugmenot1232018-10-02 09:45:22 +00:00Commented Oct 2, 2018 at 9:45
Add a comment
|
5 Answers
Just index using you ind_pos
ind_pos = [1,5,7]
print (a[ind_pos])
[88 85 16]
In [55]: a = [0,88,26,3,48,85,65,16,97,83,91]
In [56]: import numpy as np
In [57]: arr = np.array(a)
In [58]: ind_pos = [1,5,7]
In [59]: arr[ind_pos]
Out[59]: array([88, 85, 16])
answered Aug 8, 2014 at 10:32
Padraic Cunningham
181k30 gold badges264 silver badges327 bronze badges
2 Comments
MoTSCHIGGE
Ah it has to be a numpy array. If I try it like this, it doesn´t work: a = [0,88,26,3,48,85,65,16,97,83,91] ind_pos = [1,5,7] a[ind_pos] Traceback (most recent call last): File "<ipython-input-121-92d3ce287d9d>", line 1, in <module> a[ind_pos] TypeError: list indices must be integers, not list But if I add the line "arr = np.array(a)", it works. Thanks!
Padraic Cunningham
Ah ok, I presumed
a was a numpy array. Yes a would need to be a numpy arrayThe one liner "no imports" version
a = [0,88,26,3,48,85,65,16,97,83,91]
ind_pos = [1,5,7]
[ a[i] for i in ind_pos ]
Comments
Although you ask about numpy arrays, you can get the same behavior for regular Python lists by using operator.itemgetter.
>>> from operator import itemgetter
>>> a = [0,88,26,3,48,85,65,16,97,83,91]
>>> ind_pos = [1, 5, 7]
>>> print itemgetter(*ind_pos)(a)
(88, 85, 16)
3 Comments
Fabian Bosler
is this or numpy faster?
chepner
It depends. If you already have a NumPy array, that should be faster. If you have to construct a NumPy array from your list first, the time it takes to do that and the selection may be slower than it would be to simply operate on the list.
Joey
There really should be a way to do this without having to call in a package
You can use index arrays, simply pass your ind_pos as an index argument as below:
a = np.array([0,88,26,3,48,85,65,16,97,83,91])
ind_pos = np.array([1,5,7])
print(a[ind_pos])
# [88,85,16]
Index arrays do not necessarily have to be numpy arrays, they can be also be lists or any sequence-like object (though not tuples).
Comments
your code would be
a = [0,88,26,3,48,85,65,16,97,83,91]
ind_pos = [a[1],a[5],a[7]]
print(ind_pos)
you get [88, 85, 16]
