Extract data above threshold from groups

I'm trying to extract the y-data that fall into groups (bins) on the x -axis. Ideally, I want to set any y-values that are below the 90th percentile whitin a group to NaN (i.e. I want the end result to be a 6113x1 vector). I have started the following code to bin into groups , but now I am stuck.

load('xydata.mat') %

edges = (min(pp):20:max(pp));
[N, edges, bin] = histcounts(pp, edges);

figure
gscatter(pp,cc,bin)

Answer 1

[Chunru]

load('xydata.mat') 

edges = (min(pp):20:max(pp));
[N, edges, bin] = histcounts(pp, edges);

figure
gscatter(pp,cc,bin)

%whos
y= pp;
ugroups = unique(bin);
for i=1:length(ugroups)
    idx = bin == ugroups(i);    % data index for the group
    if ugroups(i) == 0          % not assigned to any group
        y(idx) = nan;
    else
        ybin = y(idx);
        pct = prctile(ybin, 90);  % 90 percentile
        ybin(ybin>pct) = nan;
        y(idx) = ybin;
    end
end



[N1, edges, bin] = histcounts(y, edges);

figure
gscatter(pp,cc,bin)

% whos
figure
plot(1:32, N, 'r', 1:32, N1, 'b')

sum(N)

ans = 6105

sum(N1)

ans = 5498

[Elin Jacobs]

Thank you very much!