1D griddedInterpolant of a vector of query points.

Is it possible to use griddedInterpolant in a vectorized form? I have a vector function f sampled over time domain t which must be interpolated.

The following relations follow: size(t,2) = size(f,2) and size(t,1) < size(f,1).

Here is a simple example that works with interp1 and spline.

Is there a way to achieve the same outcome with griddedInterpolant?

% works!
t = linspace(0,5,100);
f = [sin(t); cos(t)];              % define some vector function
pp = interp1(t, f', 'pchip','pp'); % warning, Matlab suggests to use griddedInterpolant
% or
pp2 = spline(t, f);

% This throws an error:
F = griddedInterpolant(t, f);     % ! how to make use of this function with a vector function? !

f = sin(t)                        % now f is a scalar function
F = griddedInterpolant(t, f);     % now this line works

[Star Strider]

I get no errors (R2020a) with everything except the griddedinterpolant call (that I did not run). What do you want to do? What result do you want?

[Maverick]

Please, try to run griddedinterpolant command since it does produce an error (at least on R2017b). I must interpolate a 1D vector function (i.e. a function, that has stacked multiple values at a time instant). Variable f above is a simple example of a 2-vector function. This code works with interp1 or spline, as I showed in the example. However, I want to use griddedinterpolant and check if it is more efficient thatn those counterparts.

I know that presented syntax would work if size(f,1) was equal to 1, however, I don't know how to apply griddedInterpolant to a vector function, or whether it is even possible?

Answer 1

[Matt J]

One way:

rows=1:2;
gI= griddedInterpolant({rows,t},f);    
F=@(t)gI({rows,t});

>> F(3)

ans =

    0.1411
   -0.9897

Another way:

Fsin = griddedInterpolant(t, sin(t));    
Fcos = griddedInterpolant(t, cos(t));
F=@(t)[Fsin(t);Fcos(t)];