How to structure/sort x,y,z data for surf, mesh, etc types of plots?

I'm confused conceptually about how and why my data needs to be organized to make a 3D surface plot, when I start with 3 vectors corresponding to x,y,z positions in a 3D space. Yes, I read the documentation for surf and mesh functions and other threads on similar topics, but I'm still confused.

Currently I have a 7559x3 matrix that I called A. Each row corresponds to a single x,y,z position vector; in other words, each row of A is a single point in 3D space. Simply using surf(A) does not work. The axes scales are not even correct and don't reflect my data range. Using scatter3 does work, but I want a surface plot. Here's what I get from scatter3 command:

...

I found the follwing comment/answer on a Matlab Central thread on this topic:

Three Nx1 vectors don't define a surface alone, they define a curve. You can use plot3() for that. For a surface you need to specify connectivity in some way, either by organizing the z-data into a rectangular array for surf(), or by coming up with a triangulation for trisurf()

This is what I don't understand: how do you specify connectivity? I'm especially befuddled by the part about a rectangular array. How would the z-data be organized into a rectangular array? How can you possibly make a rectangle out of the z coordinates?? It's just a single list.

[Aristo]

Were you successful in getting what you desired ? If yes could you please share your code

Answer 1

[Talmage Jones]

I believe I had a similar problem. I have three 100x1 vectors of data: two vectors of independent variables, l (with values 20,30,40...,90) and d (10,20,30,40,50), and a dependent variable, V. Reading over the documentation for the griddata function, I was able to create a matrix from this data. See script below

figure(1)

plot3(l,d,V,'.')

figure(2)

[xq,yq] = meshgrid(10:10:90, 10:10:50);

vq = griddata(l,d,V,xq,yq);

mesh(xq,yq,vq)

figure(3)

surf(xq,yq,vq)

Answer 2

[Star Strider]

It depends on how your data are organized. If you have your data on a relatively regular grid (regularly repeating values in your x and y vectors), see if griddedInterpolant does what you want. If they are not gridded but scattered, see if scatteredInterpolant works. There are a number of other interpolation functions and options that could also work, and are linked to in those pages.