How to generate a loop using logical indexes intervals?

Hi all,

I generated an index based on the depth position in which one sensor changes the direction at the water column, from upward to downward (Code below). My idea is to apply this index to locate the position of the MAXIMUM density between each index. For example, I would like to know the position of MAXIMUM density from index 1:ind(1), ind(1):ind(2), ind(2):ind(3)... The index interval are not regular and I am processing a 400810x1 size data. The figure below illustrate what I am planning to obtain, the values in red are the maximum differences in each group and I am interested to obtain the "Position" of the same. The position will work as index to continue the data processing.

Thank you for your time,

%Locating the position of change in direction based on depth differences
ind_change_direc=nan(size(depth_diff));
for i=1:length(depth_diff)-1
  if depth_diff(i)<0 & depth_diff(i+1)>0;
     ind_change_direc(i)=-9999;
  end
end
index=ind_change_direc==-9999;

Answer 1

[Star Strider]

Another possibility:

index = [-9999; NaN(7,1); -9999; NaN; NaN; -9999];
density = rand(12,1)*0.01;                          % Create Density
position = [1:12]';

change_idx = find(~isnan(index));
for k1 = 1:length(change_idx)-1
    [mx,idx] = max(density(change_idx(k1):change_idx(k1+1)-1));
    max_density(k1,:) = [mx, position(idx+change_idx(k1)-1)];
end

The ‘max_density’ matrix gives the maximum in the interval in the first column, and the value of ‘position’ for it in the second.

[Gustavo Oliveira]

Thank you very much Strider, worked greatly. However, to reach the final result I need to place the location at a size(position) matrix. The table below exemplify it at the column "Final Result", considering the same values previously explained by you. I am goint to use this "Final Result" as an index later. Probably it is a silly doubt, but I just solved the problem doing two loops, what took too long considering the size of my original data.

Kind Regards,

[Gustavo Oliveira]

The Code Below worked well to obtain the position that I was expecting ("Final Result").

Thank you again,

P=1:length(diff_dens);
P=P';
A=max_density(:,2);
for i=A(1:end)
  P(i)=-9999;
end

[Star Strider]

My pleasure.

Answer 2

[Azzi Abdelmalek]

s=[1 0  2 3 4 3 2 1 0 5 7 8 9 4 3 0 1] % Example
idx1=sign([diff(s) ])
idx1=idx1([1 1:end])
position=find(diff(idx1)~=0)+1

[Gustavo Oliveira]

Thank you for your answer Azzi. However the main intention of the routine is to identify the maximum value in each group market by the indexes (index=ind_change_direc==-9999). I did not see how the command that you wrote can work on this. I am thinking in execute a for loop using the index interval and selecting the maximum value of density, but I do not really know how.

Regards