Python numpy index

Question

i have a numpy array p like this:

array([[ 0.92691702,  0.07308298],
   [ 0.65515095,  0.34484905],
   [ 0.32526151,  0.67473849],
   ..., 
   [ 0.34171992,  0.65828008],
   [ 0.77521514,  0.22478486],
   [ 0.96430103,  0.03569897]])

If i do x=p[:,1:2], i would get

array([[ 0.07308298], [ 0.34484905], [ 0.67473849], ..., [ 0.65828008], [ 0.22478486], [ 0.03569897]]) and x.shape is (5500,1)

However, if i do x=p[:,1], i would get

array([ 0.07308298,  0.34484905,  0.67473849, ...,  0.65828008,
    0.22478486,  0.03569897])

and x.shape is (5500, )

Why there is difference like this? It quite confuses me. Thanks all in advance for your help.

Same thing with lists: [1,2,3][1] is not the same as [1,2,3][1:2]. Try understanding this example first. — shx2
– shx2, Commented Jun 13, 2014 at 18:58

mgilson · Accepted Answer · 2014-06-13 19:38:51Z

2

It's the difference between using a slice and a single integer in the ndarray.__getitem__ call. Slicing causes the ndarray to return "views" while integers cause the ndarray values.

I'm being a little loose in my terminology here -- Really, for your case they both return a numpy view -- It's easier to consider just the 1D case first:

>>> import numpy as np
>>> x = np.arange(10)
>>> x[1]
1
>>> x[1:2]
array([1])

This idea extends to multiple dimensions nicely -- If you pass a slice for a particular axis, you'll get "array-like" values along that axis. If you pass a scalar for a particular axis, you'll get scalars along that axis in the result.

Note that the 1D case really isn't any different from how a standard python list behaves:

>>> x = [1, 2, 3, 4]
>>> x[1]
2
>>> x[1:2]
[2]

edited Jun 13, 2014 at 19:38

answered Jun 13, 2014 at 18:58

mgilson

312k70 gold badges656 silver badges722 bronze badges

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

3 Comments

user2357112 Over a year ago

It's not the view/value distinction that's important here; it's the item/subsequence distinction. In fact, both operations return views for this case.

mgilson Over a year ago

@user2357112 -- Of course you're right. I was trying to draw an analogy with the 1D case, but I suppose I wasn't explicit enough. Updated.

nietzschetmh Over a year ago

Thank you all, your answers really help, i understand that now. If i use slicing, i will get a sub sequence, if i use integer, i will only get the item.

Collectives™ on Stack Overflow

Python numpy index

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related