Data filtering(Give a constraint to the length of each index )

I have this kind of data. First column and second column refer to an index and x value relatively.

data = [ 1 201; 1 202; 2 301; 2 313; 2 311; 3 401; 3 452; 3 433; 3 405; 4 504; 4 303; 4 604; 4 703; 5 600; 5 700; 5 606; 5 703; 5 905; 5 444;];

For example, I want to delete from after 4th data in each index if the number of index is over 4.

The length of 3rd, 4th and 5th index is 4, 4, 6 relatively in this example data. I'd like to keep their data only from 1st to 3rd values.

Every index has less number than 4.

data_filtered = [1 201; 1 202; 2 301; 2 313; 2 311; 3 401; 3 452; 3 433; 4 504; 4 303; 4 604; 5 600; 5 700; 5 606;];

If anyone can help, it would be greatly appreciated.

Thank you!

[Jan]

In other words: You want to keep only the first n-1 rows for each value in the first column.

Answer 1

[Jan]

There are more efficient ways, but starting with a simple loop is a good apporach:

data  = [ 1 201; 1 202; 2 301; 2 313; 2 311; 3 401; 3 452; 3 433; ...
    3 405; 4 504; 4 303; 4 604; 4 703; 5 600; 5 700; 5 606; 5 703; ...
    5 905; 5 444];
nRow  = size(data, 1);
keep  = false(1, nRow);
index = -1;  % any not occurring index
n     = 4;   % keep n-1 indices
for k = 1:nRow
    if data(k, 1) ~= index
      index = data(k, 1);
      count = 0;
    end
    count   = count + 1;
    keep(k) = count < n;
end
data_filtered = data(keep, :)

A vectorized method:

index = data(:, 1);

D     = [false ; diff(index(:)) ~= 0] ;
Start = [1; find(D)];   % Start indices of next run
N     = ones(size(D));
N(D)  = 1 - diff(Start);
N     = cumsum(N);

data_filtered = data(N < 4, :)

See also Jos' FileExchange submission https://www.mathworks.com/matlabcentral/fileexchange/56131-runindex . Then:

N = runindex(data(:, 1));
data_filtered = data(N < 4, :)