Python/Numpy: Setting values to index ranges

Question

Using Numpy, I can create a 5-dimensional array like this:

>>> faces = numpy.zeros((3, 3, 3, 6, 3))

I want (all indexes, all indexes, 0, 4) to be set to (1., 1., 1.). Is this possible by only using Numpy (no Python loops)?

NPE · Accepted Answer · 2013-12-28 09:10:51Z

6

Both of the following will do it:

faces[:,:,0,4] = 1

faces[:,:,0,4] = (1, 1, 1)

The first uses the fact that all three values are the same by having NumPy broadcast the 1 to the correct dimensions.

The second is a little bit more general in that you can assign different values to the three elements.

answered Dec 28, 2013 at 9:10

NPE

503k114 gold badges970 silver badges1k bronze badges

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

2 Comments

Is there a way I can put the slice notation in a tuple to pass to Numpy somehow?

@Synthead, You can use slice(None), slice(None), 0, 4

Andrew Walker · Accepted Answer · 2013-12-28 09:10:44Z

2

Numpy slice notation supports doing this directly

f[:,:,0,4] = 1.0

answered Dec 28, 2013 at 9:10

Andrew Walker

42.7k9 gold badges64 silver badges86 bronze badges