Piecewise constant or 0-degree spline interpolation in Python/Scipy

Question

Despite scipy's documentation indicating that scipy.interpolate.UnivariateSpline will run on order k<=5, under the hood there is an additional constraint for order >=1. Does anyone know of a way to achieve either 0-degree spline or piecewise constant interpolation with Numpy/Scipy?

Isn't what you want here? http://docs.scipy.org/doc/scipy/reference/interpolate.html — jcoppens, Jun 27 '15 at 04:02

score 3 · Answer 1 · edited Jun 27 '15 at 20:10

3

You can use interp1d with kind='zero' or kind='nearest'.

edited Jun 27 '15 at 20:10

Daniel

42,087
4
55
81

answered Jun 27 '15 at 19:44

ev-br

24,968
9
65
78

score -1 · Answer 2 · answered Aug 19 '15 at 10:40

-1

I have tried a simple trick 1) do numerical integration of step-wise constant function and you will get broken line 2) use 1. degree spline fit 3) calculate a derivative of the spline (method .derivative())

it has worked, but it could be better.

answered Aug 19 '15 at 10:40

Tomas

91
2

Piecewise constant or 0-degree spline interpolation in Python/Scipy

2 Answers2