how to save values in a matrix from a loop

Hello, im programming RLE on matlab and i'm still having one issue only and it is to save my values into a matrix instead of a vector so that when i decode, i can go row by row

this is my code

clear all;
x=[2 2 2 3 3 5; 6 6 5 5 5 7]

x = 2×6

2 2 2 3 3 5 6 6 5 5 5 7

f=size(x);
c=1;
y=[];

for i=1:f(1)
    for j=1:f(2)-1
            if x(i,j)==x(i,j+1)
            c = c + 1;
            else
            y=[y c x(i,j)];
            c=1;
            end
        
    end
    y =[y c x(i,f(2))]
        end

y = 1×6

3 2 2 3 1 5

y = 1×12

3 2 2 3 1 5 2 6 3 5 1 7

    

y               

y = 1×12

3 2 2 3 1 5 2 6 3 5 1 7

My output is coming this way y = 3 2 2 3 1 5 2 6 3 5 1 7 and instead i want it something like y = 3 2 2 3 1 5

2 6 3 5 1 7

[Voss]

What would you do if the number of unique values on each row of x is not the same? For example, if

x = [2 2 2 3 3 5; 6 6 5 5 5 5]

then what would y be? Maybe this?:

y = [3 2 2 3 1 5; 2 6 4 5 0 0]

Answer 1

[Voss]

To handle the case that not all rows of x have the same number of unique values, you can make y a cell array and construct each cell the same as you are already doing:

clear all;
x=[2 2 2 3 3 5; 6 6 5 5 5 5]

x = 2×6

2 2 2 3 3 5 6 6 5 5 5 5

f=size(x);
c=1;
y=cell(1,f(1)); % make a cell in y for each row of x
y

y = 1×2 cell array
    {0×0 double}    {0×0 double}

for i=1:f(1)
    for j=1:f(2)-1
        if x(i,j)==x(i,j+1)
            c = c + 1;
        else
            y{i} = [y{i} c x(i,j)];
            c=1;
        end
    end
    y{i} = [y{i} c x(i,f(2))];
end
y

y = 1×2 cell array
    {[3 2 2 3 1 5]}    {[2 6 4 5]}

If all cells of y contain vectors of the same length, as in your example x, then you can make y a matrix like this:

new_y = cell2mat(y(:)) % won't work if not all cells of y are the same length

If not, then you can make a matrix of zeros (or NaNs or whatever) and fill in each row to the appropriate column:

N = max(cellfun(@numel,y));
new_y = zeros(f(1),N);
for i = 1:f(1)
    new_y(i,1:numel(y{i})) = y{i};
end
new_y

new_y = 2×6

3 2 2 3 1 5 2 6 4 5 0 0

[Adel Hafri]

Cells are totally what i was looking for, still new to coding so didnt know about them before

thank you so much!

[Voss]

You're welcome!

[Jan]

A small simplification:

sx = size(x);
y  = cell(1, sx(1));  % pre-allocate the output
for i = 1:sx(1)       % Loop over rows
    c = 1;
    a = [];           % Accumulate
    for j = 1:sx(2) - 1
        if x(i, j) == x(i, j+1)
            c = c + 1;
        else
            a = [a, c, x(i, j)];  % append to the last cell of y
            c = 1;
        end
    end
    y{i} = [a, c, x(i, sx(2))];   % Insert in output
end

Answer 2

[Walter Roberson]

clear all;
x=[2 2 2 3 3 5; 6 6 5 5 5 7]

x = 2×6

2 2 2 3 3 5 6 6 5 5 5 7

f=size(x);
c=1;
y = {};

for i=1:f(1)
    y{i} = [];
    for j=1:f(2)-1
            if x(i,j)==x(i,j+1)
            c = c + 1;
            else
            y{i}=[y{i} c x(i,j)];
            c=1;
            end
        
    end
    y{i} =[y{i} c x(i,f(2))]
end

y = 1×1 cell array
    {[3 2 2 3 1 5]}

y = 1×2 cell array
    {[3 2 2 3 1 5]}    {[2 6 3 5 1 7]}

y               

y = 1×2 cell array
    {[3 2 2 3 1 5]}    {[2 6 3 5 1 7]}

However you should not proceed to vertcat(y{:}) because you have no reason ahead of time to believe that the encodings will be the same length for each row.

Answer 3

[Jan]

A completely different approach:

x    = [2 2 2 3 3 5; 6 6 5 5 5 7];
nRow = size(x, 2);
y    = cell(1, nRow);  % Pre-allocate: avoid iterative growing of arrays
for k = 1:nRow
  [b, n] = RunLengthEnc(x(k, :));   % Get elements and count them
  y{k}   = reshape([n; b], 1, []);  % Store a row vector
end

% Determine run length parameters without a loop:
function [b, n] = RunLengthEnc(x)
d = [true, diff(x) ~= 0];   % TRUE if values change
b = x(d);                   % Elements without repetitions
n = diff(find([d, true]));  % Number of repetitions
end