[Numpy-discussion] Questions on np.piecewise

[Eric Firing]

On 2017/09/15 2:02 AM, Joe wrote:
> Hello,
> 
> I have two questions and hope that you can help me.
> 
> 1.)
> Is np.piecewise only defined for two conditions or should something like
> 
> [0 < x <= 90, 90 < x <= 180, 180 < x <= 270]
> 
> also work?
> 
> 2.)
> Why does
> 
> np.piecewise(np.array([50]), [0 < x <= 90, 90 < x <= 180], [1.1, 2.1])
> 
> return [2] and not [2.1] ?
> 
> Kind regards,
> Joe

Your example doesn't run, but here is one that does:

In [8]: x = np.array([50], dtype=float)

In [9]: np.piecewise(x, [0 < x <= 90, 90 < x <= 180], [1.1, 2.1])
array([ 1.1])

The answer to your second question is that it is returning an array with 
the same dtype as its first argument.

The answer to your first question is "yes", and evidently if more than 
one condition matches, it is the last that prevails:

In [10]: np.piecewise(x, [0 < x <= 90, 90 < x <= 180, 30<x<80], [1.1, 
2.1, 3.3])
array([ 3.3])

Eric