Spline with clamped end conditions using griddedInterpolant

I can create a piecewise polynomial spline with clamped end conditions from vectors x and y using

pp = spline(x,[dy1 y dy2])

where dy1 and dy2 are the required slopes. How can I do this using griddedInterpolant?

fy = griddedInterpolant(x,y,'spline')

is the knot-a-knot spline. Can I not clamp the end points?

Answer 1

[John D'Errico]

You cannot do so. There is no capability for that in griddedInterpolant. But I fail to see the problem. Just use spline, which DOES allow that option. Then you can interpolate the spline you have created as you wish. WTP?

[John Billingham]

Sorry, I should have added that griddedinterpolant seems to be much faster than ppval, and it's this speed that I'm after for a clamped spline.

[John Billingham]

x = linspace(0,1,100); y = rand(1,100);
x1 = rand(1,100);
ppx = spline(x,y);
fx = griddedInterpolant(x,y,'spline');
tic
for k = 1:100000
    y1 = ppval(ppx,x1);
end
toc

Elapsed time is 1.429425 seconds.

tic
for k = 1:100000
    y1 = fx(x1);
end
toc

Elapsed time is 0.300864 seconds.

This makes a considerable difference for what I'm trying to do.