convert cell into double with diferent cell sizes

Hello, for example i have

A = [1,2;2,4;3,NaN;NaN,5;NaN,7]
res = cellfun(@(x) find (~isnan(x)),num2cell(A,1),'un',0) % Here i store in the index of the numbers that are not NaN

res is a 1x2 cell array, the first component is a 3x1 double and the second one is a 4x1 double.

res{1} = 1
         2
         3
res{2} = 1
         2
         4
         5

My question is to do a table or a double, or a matrix that each res{j} is stored in the respective j column of the new table I tried a cell2mat, but since they aren't the same size it doesn't work. the goal is to make

output = 1 1
         2 2
         3 4  
         (empty or NaN or whatever) 5

[Bob Thompson]

I know that it's possible to pad an array, but I'm not sure of the code. I would suggest making an array with the number of columns equal to the largest cell in res, then filling in the values from the cells of res. All extra elements will be NaN.

array = NaN(max(size(res{:})),size(res));
for I = 1:size(res);
    array(:,I) = res(I);
end

I also can't guarantee that size works properly with cell arrays, but that should at least get you started.

[Tiago Dias]

i did this, but only saves me for j=m, he does for j=1, for j=2, but only saves for j=2

[n,m] = size(A);
j = 1:m
table = res{j};
end

but doesnt work, because it only saves me the values of the 2nd cell

[Tiago Dias]

[n,m] = size(A);
j = 1
table = res{j};
end
j = 2
table2 = res{j};
end
he does what i want, but my original data have 140 j xD

[Bob Thompson]

Are you actually trying to get rid of all of the NaN values, or just gather the non-NaN values together? If you're just trying to gather the values try using the following:

A = [1,2;2,4;3,NaN;NaN,5;NaN,7]
res = cellfun(@(x) find (~isnan(x)),num2cell(A,1),'un',0)
locationarray = NaN(size(A)); % Create an array the size of A, but filled with NaN values
for I = 1:size(A,2) % Run index I through each column of A, this should correspond to each cell in res
  locationarray(:,I) = res{I}; % Put cell I of res into column I of the location array. Because it's in a for loop it will cover all columns
end

[Tiago Dias]

I want to store the index on NaN values, the way I do I store that in a cell, but I want to store if possible in a double, table or matrix. just that. But my input have values and NaN values numbers.

The code you wrote gets me an error:

Subscripted assigment dimension mismatch 
Error in (line 5)
locationarray(:,I) = res{I};

If i use

locationarray = res{I}
It only store in location array the information when I=2, doesnt save for I=1

Answer 1

[Jos (10584)]

You can my PADCAT function to concatenate vectors with different lengths. It is available at the Matlab File Exchange:

https://uk.mathworks.com/matlabcentral/fileexchange/22909-padcat-varargin-

However, there are simpler ways to get the output res, from A:

A = [1,2;2,4;3,NaN;NaN,5;NaN,7]

res = repmat(1:size(A,1),size(A,2),1)' ;
res(isnan(A)) = NaN
res = sort(res)

Answer 2

[Tiago Dias]

Thanks, works perfect! Will try on my big data to see if is good