46

I am working in Python and I have a NumPy array like this:

[1,5,9]
[2,7,3]
[8,4,6]

How do I stretch it to something like the following?

[1,1,5,5,9,9]
[1,1,5,5,9,9]
[2,2,7,7,3,3]
[2,2,7,7,3,3]
[8,8,4,4,6,6]
[8,8,4,4,6,6]

These are just some example arrays, I will actually be resizing several sizes of arrays, not just these.

I'm new at this, and I just can't seem to wrap my head around what I need to do.

3 Answers 3

50

@KennyTM's answer is very slick, and really works for your case but as an alternative that might offer a bit more flexibility for expanding arrays try np.repeat:

>>> a = np.array([[1, 5, 9],
              [2, 7, 3],
              [8, 4, 6]])

>>> np.repeat(a,2, axis=1)
array([[1, 1, 5, 5, 9, 9],
       [2, 2, 7, 7, 3, 3],
       [8, 8, 4, 4, 6, 6]])

So, this accomplishes repeating along one axis, to get it along multiple axes (as you might want), simply nest the np.repeat calls:

>>> np.repeat(np.repeat(a,2, axis=0), 2, axis=1)
array([[1, 1, 5, 5, 9, 9],
       [1, 1, 5, 5, 9, 9],
       [2, 2, 7, 7, 3, 3],
       [2, 2, 7, 7, 3, 3],
       [8, 8, 4, 4, 6, 6],
       [8, 8, 4, 4, 6, 6]])

You can also vary the number of repeats for any initial row or column. For example, if you wanted two repeats of each row aside from the last row:

>>> np.repeat(a, [2,2,1], axis=0)
array([[1, 5, 9],
       [1, 5, 9],
       [2, 7, 3],
       [2, 7, 3],
       [8, 4, 6]])

Here when the second argument is a list it specifies a row-wise (rows in this case because axis=0) repeats for each row.

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

Comments

19
>>> a = numpy.array([[1,5,9],[2,7,3],[8,4,6]])
>>> numpy.kron(a, [[1,1],[1,1]])
array([[1, 1, 5, 5, 9, 9],
       [1, 1, 5, 5, 9, 9],
       [2, 2, 7, 7, 3, 3],
       [2, 2, 7, 7, 3, 3],
       [8, 8, 4, 4, 6, 6],
       [8, 8, 4, 4, 6, 6]])

2 Comments

In case someone's wondering, it's the Kronecker product: docs.scipy.org/doc/numpy/reference/generated/numpy.kron.html
as cool as this answer is, it takes twice as long as the repeat method in dtlussier's answer on my machine for big arrays
2

Unfortunately numpy does not allow fractional steps (as far as I am aware). Here is a workaround. It's not as clever as Kenny's solution, but it makes use of traditional indexing:

>>> a = numpy.array([[1,5,9],[2,7,3],[8,4,6]])
>>> step = .5
>>> xstop, ystop = a.shape
>>> x = numpy.arange(0,xstop,step).astype(int)
>>> y = numpy.arange(0,ystop,step).astype(int)
>>> mg = numpy.meshgrid(x,y)
>>> b = a[mg].T
>>> b
array([[1, 1, 5, 5, 9, 9],
       [1, 1, 5, 5, 9, 9],
       [2, 2, 7, 7, 3, 3],
       [2, 2, 7, 7, 3, 3],
       [8, 8, 4, 4, 6, 6],
       [8, 8, 4, 4, 6, 6]])

(dtlussier's solution is better)

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.