how does griddata asses duplicates?

Hi! I would like to understand how griddata asses duplicate point. I opened the function and I found this section:

%Need x,y and z to be column vectors
sz = numel(x);
x = reshape(x,sz,1);
y = reshape(y,sz,1);
v = reshape(v,sz,1);
myepsx = eps(0.5 * (max(x) - min(x)))^(1/3);
myepsy = eps(0.5 * (max(y) - min(y)))^(1/3);


% look for x, y points that are indentical (within a tolerance)
% average out the values for these points
if isreal(v)
    xyv = builtin('_mergesimpts', [y, x, v], [myepsy, myepsx, Inf], 'average');
    x = xyv(:,2);
    y = xyv(:,1);
    v = xyv(:,3);  

could somebody explain me

xyv = builtin('_mergesimpts', [y, x, v], [myepsy, myepsx, Inf], 'average');

Answer 1

[Adam Danz]

That line calls Matlab's built-in static method or function matlab.internal.math.mergesimpts(). The comments describe what it's doing. It takes the x y v vectors and defines tolerance levels for each variable (myepsy, myepsx, Inf). The tolerance levels define which values are "identical" or "close enough to be identical". Identical values are then averaged to get rid of (near) duplicates.