Is there a faster way to concantenate this matrix?

I have a matrix A that is initially a 2 x 2 matrix.

I compute a new value of A and call this A_new which is also a 2 x 2 matrix

I concatenate the matrix by A = [A A_new] and store it back into A.

I compute another value and call it A_new2 and continue this. Essentially, I am doing the following :

A = [A A_new A_new2 A_new3 ....]

I think this is slowing me down since the size of A grows. Is there a better way to do this?

Answer 1

[Azzi Abdelmalek]

Pre-allocate

A=zeros(2,1000)
A(:,1:2)=A1;
A(:,3:4)=A2;
%and so on

[John]

Thanks but the number of A1, A2, A3 are unknown. If I have 1000 As, then it wouldn't be efficient.

[Jan]

@Azzi: Did you edit your answer after John's comment? If so, you see how confusing his comment gets by this. Please avoid such confusions by adding new code instead of changing the existing without a notice. If you didn't, I'm confused by John's comment and the accepting.

[Azzi Abdelmalek]

Jan, I don't remember that, but if I did it, The forum editor should mark it as edited

Answer 2

[Matt J]

 C{1}=A
 C{2}=A_new

 etc...

Then, at the end, you would do

 A=horzcat(C{:});

[Matt J]

It is better, in this approach to pre-allocate C

 C=cell(1,N);

Even if you do not know precisely what N will be, it is better to pre-allocate with an over-estimated N, than nothing at all.