find max of every 4 rows in a matrix and save in separate array

Im trying to find max of 5th column of every 4 rows in a matrix say A. I need to push the max rows into a separate matrix B and remove these rows from A. I have attached sample data herewith. The data is for 6 customers with 4 tags(12,22,32,42) for each customer. I need to take the max of percentage between the four tags for each customer. So far I have tried with a for loop as :

custs = unique(data(:,3));
highertrp = zeros(size(custs,1),1);

for i = 1 : size(custs,1)
    idx = find(data(:,3)== custs(i,1));
    [~,id] = max(data(idx,5),[],1);
    highertrp(i,1:size(data,2)) = data(idx(id),:);
    data(idx(id),:) = [];
end

But this is too slow as I can have about 1 million records to check in my project. Any help will be great.

Answer 1

[Andrei Bobrov]

data  = xlsread('sample_data.xlsx',1);
[~,~,c] = unique(data(:,3));
idx =  accumarray(c,(1:size(data,1))',[],   @(x)  x(  max(data(x,5))==data(x,5) ,:) );
highertrp = data(idx,:);
data(idx,:) = [];

in case if find max of every 4 rows

data  = xlsread('sample_data.xlsx',1);
[~,ii] = max(reshape(data(:,5),4,[]));
idx = ii + (0:numel(ii)-1)*4;
highertrp = data(idx,:);
data(idx,:) = [];

[Hariprasad]

Thanks for the reply. What is 'c' in

accumarray(c,(1:size(data,1))',[],  @(x)  x(  max(data(x,5))==data(x,5) ,:) );

Anyway, the second method worked.I should have thought about reshape().

Thanks a ton @Andrei Bobrov

[Andrei Bobrov]

Sorry, it my typo. Corrected.

[Hariprasad]

Oh ok the indices of each Id. Thank you @Andrei Bobrov