Normalize matrices of different sizes

I have different size matrix a1, a2, a3...an, size of a1= [m n1]; size of a2=[m n2]; size of a3=[m n3];...size of an=[m nn]; this means that the row number of all the matrix is the same, but their columns number is different. How could I normalize them into matrix [m n], while n is the biggest column number of all matrix? the rest of colums of other matrix should be added with zero. finally, how can I save all new matrix into a 3-D matrix?

Answer 1

[Ameer Hamza]

See this example

a1 = rand(3,4);
a2 = rand(3,7);
a3 = rand(3,2);
a4 = rand(3,5);

a = {a1, a2, a3, a4};
m = height(a{1});
n = max(cellfun(@width, a));
a = cellfun(@(x) {[x zeros(m, n-width(x))]}, a);

A = cat(3, a{:}) % 3D matrix

[xiao wei]

thanks for your reply.

now there is another problem: when run on this sentence

'a = cellfun(@(x) {[x zeros(m, n-width(x))]}, a);'

there is something wrong, the note is: can not convert from 'double' to 'struct'.

how could I do in this case?

[Ameer Hamza]

Can you show your code?

[xiao wei]

my code is as follows:

loadstr(:,1) ={'...'}; %file name

loadstr(:,2) = {... }; % file classification

for i = 1:numel(loadstr(:,1))

path = 'C:\...'; % path for file

file = loadstr{i,1};

fileName = strcat(path,file,'.mat');

vars{i} = load(fileName);

tmp (i) = structfun(@length,vars{i});

maxLength=max(tmp);

end

for i = 1:numel(loadstr(:,1))

a = cellfun(@(x) {[x zeros(126, maxLength-tmp(i))]}, vars(i));

end

my problem is:

1. how can I add zeros(126, maxLength-tmp(i)) to the end of each file and save it as new file (in the name of old file)?
2. how can I save these new files into a 3-D matrix?