Split numpy array at multiple values?

Question

Based on my question Fastest way to approximately compare values in large numpy arrays? I was looking for ways to split an array as I wanted. I have a sorted array (2D, sorted by values in one column), and want to split it into multiple arrays. Not of equal length based on index, but of equal range in values. The closest question I found is Split array at value in numpy but I'd like to do something a bit different. Say I have (1D example):

[0.1, 3.5, 6.5, 7.9, 11.4, 12.0, 22.3, 24.5, 26.7, 29.9]

and I want to split it into ranges [0,10) [10,20) [20,30] so it becomes

[0.1, 3.5, 6.5, 7.9] [11.4, 12.0] [22.3, 24.5, 26.7, 29.9]

Bi Rico · Accepted Answer · 2012-08-01 20:48:05Z

7

The 1d case can be done like this

>>> A = np.array([0.1, 3.5, 6.5, 7.9, 11.4, 12.0, 22.3, 24.5, 26.7, 29.9])
>>> split_at = A.searchsorted([10, 20])
>>> B = numpy.split(A, split_at)

This also works in 2d, if I understood your question correctly, for example:

>>> A = array([[  0.1,   0. ],
               [  3.5,   1. ],
               [  6.5,   2. ],
               [  7.9,   3. ],
               [ 11.4,   4. ],
               [ 12. ,   5. ],
               [ 22.3,   6. ],
               [ 24.5,   7. ],
               [ 26.7,   8. ],
               [ 29.9,   9. ]])
>>> split_at = A[:, 0].searchsorted([10, 20])
>>> B = numpy.split(A, split_at)
>>> B
[array([[ 0.1,  0. ],
       [ 3.5,  1. ],
       [ 6.5,  2. ],
       [ 7.9,  3. ]]),
 array([[ 11.4,   4. ],
       [ 12. ,   5. ]]),
 array([[ 22.3,   6. ],
       [ 24.5,   7. ],
       [ 26.7,   8. ],
       [ 29.9,   9. ]])]

edited Aug 1, 2012 at 20:48

answered Aug 1, 2012 at 20:37

Bi Rico

25.9k3 gold badges57 silver badges75 bronze badges

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

1 Comment

Tristan Klassen Over a year ago

For a 2D array, it looks like I slice the appropriate column and then use searchsorted on that.

Collectives™ on Stack Overflow

Split numpy array at multiple values?

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related