迭代函数输出到矩阵中

Question

I have an iterative function which gives me two vector outputs. How can I store these outputs into two separate matrices in Matlab?

[A, B]=iterative_function(x,y)

The size of A and B does not vary, they are 1x4.

Is there a way to do this without 'cell'?

Answer 1

If I understood correctly you want to concatenate all the A 's and B 's that are output at every iteration. If their size never change you don't need to use cell arrays, you can use a regular matrix.
Your A and B are row-vectors, so let's say you want to concatenate each result as a single row.

A_total=[];
B_total=[];
while(your loop is going)
    [A, B]=iterative_function(x,y)
    A_total=[A_total ; A];
    B_total=[B_total ; B];
end

If you know a priori how many iterations will be done by your loop you can as well preallocate such matrices A_total and B_total and assign each A and B to the i -th row:

A_total=zeros(MaxIteration,4); %because A has length 4
B_total=zeros(MaxIteration,4); %because B has also length 4
for i=1:MaxIterations
    [A, B]=iterative_function(x,y)
    A_total(i,:)=A;
    B_total(i,:)=B;
end

Update : as @Max correctly pointed out, I reckon is also worth knowing that you can do the very same thing (ie concatenate several vectors of equal size in a matrix) even if your vectors are not row-vectors but column-vectors. The main trick is that now you do not assign 1 row = 1 vector, but 1 column = 1 vector. Let's suppose that A and B are not 1x4, but 4x1. Such concatenation can easily be done as follows (inside the loop):

    A_total=[A_total , A]; % a space without a comma works just as well
    B_total=[B_total , B];

Or, again, if you know a priori the number of iteration you can preallocate in this manner:

A_total=zeros(4,MaxIteration); %because A has length 4
B_total=zeros(4,MaxIteration); %because B has also length 4

this time you'll have as many columns as there are iterations whereas in the previous case you had as many rows as there were iterations. And inside the loop you can replace the i-th column in this manner:

    A_total(:,i)=A;
    B_total(:,i)=B;

where (my bad I didn't explain earlier) the colon operator ( : ) means "all of them" (ie the expression A_total(:,i) means the i-th column and all the rows).

Finally, both of these methods work because A and B have the same size and it'll never change. Although if, for any reasons, such sizes turn out to be different, Matlab will not allow you to concatenate such vectors and your only chance relies on cell arrays. A matrix, as you'd expect, can only be created if all vectors have the same size (ie you cannot concatenate an array of length 4 and an array of length 5 in two rows). Contrarily, a cell array is an heterogeneous data structure so each cell can contain whatever data you like.