Numpy array documentation slicing rule

Question

In Basic slicing of numpy array http://docs.scipy.org/doc/numpy-1.5.x/reference/arrays.indexing.html#basic-slicing,
I found the following rule which does not work for the example I shown below.

Rule: Assume n is the number of elements in the dimension being sliced. Then, if i is not given it defaults to 0 for k > 0 and n for k < 0 . If j is not given it defaults to n for k > 0 and -1 for k < 0 . If k is not given it defaults to 1. Note that :: is the same as : and means select all indices along this axis.

What I understood : This can be prioritized as top to down :

a)If k is not given it defaults to 1.
b)If j is not given it defaults to n for k > 0 and -1 for k < 0 .
c)if i is not given it defaults to 0 for k > 0 and n for k < 0.

Now let's see the example. Basically what I am doing is to take a 3d array and print it bottom up . Layer with largest index comes first and then the smaller one. Please look at code for better understanding.

import numpy as np

b= np.arange(24).reshape(2,3,4)
print "Here is the input :" 
print b
print

print "Here is what is desired output :"
print b[::-1 , :: ,::]
print 

print "Here is what I want to get desired output by a different way using above rule :"
print b[2:-1:-1 , :: , ::]

Output :

Here is the input :
[[[ 0  1  2  3]
  [ 4  5  6  7]
  [ 8  9 10 11]]

 [[12 13 14 15]
  [16 17 18 19]
  [20 21 22 23]]]

Here is what is desired output :
[[[12 13 14 15]
  [16 17 18 19]
  [20 21 22 23]]

 [[ 0  1  2  3]
  [ 4  5  6  7]
  [ 8  9 10 11]]]

Here is what I want to get desired output by a different way using above rule :
[]

Is b[::-1 , :: ,::] not same as b[2:-1:-1 , :: , ::] by above rule?

HYRY · Accepted Answer · 2016-01-29 08:41:12Z

2

The document is right, but it doesn't means that you can use the calculated start, end index in slice object again. It only tells you the logic to calculate the start & end index. To use the calculated index, you need to generate the index by range():

Here is an example:

import numpy as np
s = slice(None, None, -1)
t = np.array([1, 2, 3, 4])
s.indices(len(t))

outputs:

(3, -1, -1)

so the (start, stop, stride) of [::-1] for a four element array is (3, -1, -1), but t[3:-1:-1] is empty. The (start, stop, stride) is for range(), So, you can use t[range(3,-1,-1)] which is [4, 3, 2, 1].

edited Jan 29, 2016 at 8:41

answered Jan 29, 2016 at 8:33

HYRY

97.8k28 gold badges197 silver badges192 bronze badges

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

2 Comments

Number945 Over a year ago

Slice objects are also generated when extended indexing syntax is used. For example: a[start:stop:step] or a[start:stop, i] . Pls have a look at : docs.python.org/dev/library/functions.html#slice ; then why t[3:-1:-1] gives empty ?

HYRY Over a year ago

t[-3:-1:-1] gives empty because slice(-3, -1, -1).indices(4) returns (1, 3, -1).

hpaulj · Accepted Answer · 2016-01-29 08:38:53Z

With a negative step, you need to think twice about the start and stop values.

A simple reverse:

In [437]: np.arange(10)[::-1]
Out[437]: array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])

start at the end - same thing

In [438]: np.arange(10)[10::-1]
Out[438]: array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])

with -1 stop, empty

In [439]: np.arange(10)[10:-1:-1]
Out[439]: array([], dtype=int32)

but with -2, it returns the last value; remember it's counting from the end, and -2 also counts from the end.

In [440]: np.arange(10)[10:-2:-1]
Out[440]: array([9])

last 4 items:

In [441]: np.arange(10)[10:-5:-1]
Out[441]: array([9, 8, 7, 6])

all but the first:

In [442]: np.arange(10)[10:-10:-1]
Out[442]: array([9, 8, 7, 6, 5, 4, 3, 2, 1])

same with 0:

In [445]: np.arange(10)[10:0:-1]
Out[445]: array([9, 8, 7, 6, 5, 4, 3, 2, 1])

you have to use None to get everything:

In [446]: np.arange(10)[10:None:-1]
Out[446]: array([9, 8, 7, 6, 5, 4, 3, 2, 1, 0])

Collectives™ on Stack Overflow

Numpy array documentation slicing rule

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related