How to plot a 3D surface with two vectors and one array?

I want to plot a 3D surface , but I cannot apply the surf() command straightforwardly for some reason.

I have two vectors for x and y coordinates. Let us say

x = linspace(-10, 10, 100);
y = linspace(-0.5, 0.5, 100);

To calzulate I have to call some function which works only for fixed values of x and y. Thys, I calculate it as follows

for i = 1 : length(x)
    for j = 1 : length(y)
        func(i,j) = @MyFunction(__some_parameters__, x(i), y(j));
    end
end

As a result I have two vectors for x and y and a matrix of the size length(x)*length(y) for the variable z. Could you tell me please, if there are any convenient methods to rewrite this data to a format that is appropriate for applying the surf() command?

P.S. I used plot3(x(i), y(j), func(i,j), 'b.') with two cycles, but it is a ridiculous "dirty hack" and not a surface.

Answer 1

[Walter Roberson]

Create full vectors of coordinates and arrayfun()

x = linspace(-10, 10, 100);
y = linspace(-0.5, 0.5, 100);
[X, Y] = meshgrid(x, y);
Z = arrayfun(@(xx,yy) sin(xx)+atan2(yy,xx), X, Y);
surf(x, y, Z, 'edgecolor', 'none')

[Bogdan MP]

Tanhk you. But, unfortunately, it does not applicable to my case (or, at least, I have no idea how to write it correctly). The function I have to call, in fact has the following structure

MyFunc(@OtherFunc, p(1), p(2), .. , p(n), x, y)

here OtherFunc - is some another function of the parameters and . And are just some fixed numbers. This function blows up for some parameters and I accurately control it in a cycle.

Is it possible to reorganise the data itself? Without recalling inner functions? Let's say the only data I have is and an array of .

[Bogdan MP]

Well, it looks like I can simply write

surf(x, y, z')

[Walter Roberson]

surf(x, y, z')

If that worked for your purposes, then chances are that when you constructed z, you varied x values down columns, and that you varied y values across rows -- that array z(J,K) is a function of x(J), y(K) . That is a very common data organization... but it is not the MATLAB data organization. In MATLAB, up/down (so, along columns) is y, and left/right (so, along rows) is x, and array location z(J,K) is z(y(J), x(K)) .

When you use meshgrid(xvector, yvector), then the coordinates are arranged in such a way that surf(x, y, z) works out.

When you use ndgrid(xvector, yvector) then the coordinates are arranged in such a way that surf(x, y, z) errors (unless xvector and yvector are the same length).

xvec = 1:2; yvec = 1:3

yvec = 1×3

1 2 3

[X, Y] = meshgrid(xvec, yvec)

X = 3×2

1 2 1 2 1 2

Y = 3×2

1 1 2 2 3 3

[X, Y] = ndgrid(xvec, yvec)

X = 2×3

1 1 1 2 2 2

Y = 2×3

1 2 3 1 2 3