Average data along z

I have data at given xyz points (i.e. locations). For each xy pair, there are multiple z values. I'd like to average the data along the z dimension. In other words, go from 3D+Value to 2D+AveragedValueAlongZ. Is there a fast way to do this (i.e. not needing loops)? Thanks in advance.

Answer 1

[Walter Roberson]

[uxy, ~, uidx] = unique([x(:), y(:)], 'rows');
avgv = accumarray(uidx, v(:), [], @mean);
output = [uxy, avgv];

[Walter Roberson]

The above can be used for scattered points where there are not necessarily the same number of z for each x/y combination.

[Alireza H.]

Thank you so much, Walter! This is exactly what I needed. My data was actually 4D: [ x y z Value] and I wanted 3D z-averaged: [x y AveragedValueAlongZ] but a minor change to your code did the job.

Answer 2

[jonas]

mean(A,3)

[Walter Roberson]

This can be useful, but does depend upon the points being gridded, with an equal number of z points for each (x,y) pair.

[jonas]

Ah yes, that's how I interpret 3D to 2D, but in hindsight I agree that OP is probably looking for your solution.