Passing list as slice for a N-dimensional numpy array

Question

I'm trying to manipulate the values of a N-dimensional array based on the user's decision, at which index the array should be changed. This example works fine:

import numpy as np
a = np.arange(24).reshape(2,3,4)

toChange = ['0', '0', '0'] #input from user via raw_input

a[toChange] = 0

But if I want to change not only one position but a complete row, I run into problems:

toChange = ['0', '0', ':'] #input from user via raw_input

a[toChange] = 0

This causes ValueError: setting an array element with a sequence. I can see that the problem is the ':' string, because a[0, 0, :] = 0 does exactly what I want. The question is, how to pass the string to the array?

Or is there a smarter way to manipulate user-defined slices?

PS: as I'm working on an oldstable Debian I use Python 2.6.6 and Numpy 1.4.1

Fred Foo · Accepted Answer · 2014-02-18 14:39:28Z

1

: is syntactic sugar for a slice object:

>>> class Indexable(object):
...     def __getitem__(self, idx):
...         return idx
...     
>>> Indexable()[0, 0, :]
(0, 0, slice(None, None, None))

So if you replace ':' with slice(None, None, None) you get the desired result:

>>> toChange = [0, 0, slice(None, None, None)]
>>> a[toChange] = 0
>>> a
array([[[ 0,  0,  0,  0],
        [ 4,  5,  6,  7],
        [ 8,  9, 10, 11]],

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

answered Feb 18, 2014 at 14:39

Fred Foo

365k80 gold badges765 silver badges852 bronze badges

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

4 Comments

Gnihilo Over a year ago

Thank you for the quick response! This works nicely in the interactive mode, but somehow within my script I get a new error: toChange = [0, 0, slice(None, None, None)] TypeError: 'numpy.ndarray' object is not callable Is there a possibility that toChange is considered to be an array?

seberg Over a year ago

Using a tuple to do ND-indexes is preferable. List/Sequence usage depends on shaky logic and I would personally prefer to just forbid it.

Gnihilo Over a year ago

Ok, I got the problem: the slicing didn't work, because before manipulating the array I was writing the whole array to a file using

with open("out/histData.txt", "w") as dataFile:     for slice in histData:         np.savetxt(dataFile, slice, fmt="%.2f", delimiter="\t")

Therefore slice was an ndarray and the whole thing didn't work. Amateur hour :-) @seberg: how would this solution look like?

seberg Over a year ago

@Gnihilo, exactly the same, just make sure toChange is a tuple at the end (if you need the mutability if the list, just call tuple(toChange))

Collectives™ on Stack Overflow

Passing list as slice for a N-dimensional numpy array

1 Answer 1

4 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Related