[Numpy-discussion] simple question

[Mathew Yeates]

ohhhh. I was looking at using "where"



Keith Goodman wrote:
> On 7/5/06, Mathew Yeates <myeates at jpl.nasa.gov> wrote:
>> What is the typical way of doing the following
>> starting with a 0 matrix, set all values to 1 when a certain condition
>> is met, set to -1 when another condition is met, left alone if neither
>> condition is met.
>
> This works on recent versions of numpy:
>
>>> x = asmatrix(zeros((2,2)))
>
>>> x
>
> matrix([[0, 0],
>       [0, 0]])
>
>>> y = asmatrix(rand(2,2))
>
>>> y
>
> matrix([[ 0.85219404,  0.48311427],
>       [ 0.41026966,  0.2184193 ]])
>
>>> x[y > 0.5] = 1
>
>>> x[y < 0.5] = -1
>
>>> x
>
> matrix([[ 1, -1],
>       [-1, -1]])
>