Getting the first occurrence of a value in an N-dimensional numpy array

Question

I have seen this question, but want to reduce the array created from mask = array == value

mask = array([[[ True,  True,  True],
               [False,  True,  True]],

              [[False,  True,  True],
               [False,  True,  True]],

              [[False, False,  True],
               [False,  True,  True]]])

which results in

where(mask) = (array([0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2]),
               array([0, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1]),
               array([0, 1, 2, 1, 2, 1, 2, 1, 2, 2, 1, 2]))

and I want to reduce it to an array of the first occurrences of True

array([[0, 1],
       [1, 1],
       [2, 1]])

but can't work out how to go about this from the output of numpy.where. Can anyone help me out?

If your array has 3 dimensions, shouldn't the first occurrence of True be array([[0],[0],[0]])? What does "an array of the first occurrences" mean? — wflynny
– wflynny, Commented Mar 6, 2014 at 23:05
One solution could be with np.argmax possibly with an axis defined. — M4rtini
– M4rtini, Commented Mar 6, 2014 at 23:17

M4rtini · Accepted Answer · 2014-03-06 23:20:19Z

2

Actually, it's as simple as this:

np.argmax(mask, 2)

Example:

In [15]: %paste
mask = array([[[ True,  True,  True],
               [False,  True,  True]],

              [[False,  True,  True],
               [False,  True,  True]],

              [[False, False,  True],
               [False,  True,  True]]])

## -- End pasted text --

In [16]: np.argmax(mask, 2)
Out[16]:
array([[0, 1],
       [1, 1],
       [2, 1]], dtype=int64)

answered Mar 6, 2014 at 23:20

M4rtini

13.6k4 gold badges37 silver badges42 bronze badges

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

Collectives™ on Stack Overflow

Getting the first occurrence of a value in an N-dimensional numpy array

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related