Capturing numbers from different sized elements in for loop

I have a for loop with multiple commands and operators, at the end of the loop I am given important numbers that I would like to store with each iteration of the loop. Each iteration of the loop may contain a different number of values to store. In order to store the values I have used a cell array. This works but I need an easy way to access the numbers within the cell. Below is a similar, though generic, situation to the problem I am facing.

for i=1:10
    z=randi(5,1);
    store=randi(200,z,1);
    all{i}=store;
end

I would like to be able to access the numbers in "all" as doubles or matrices, as opposed to cells or strings. Any help is appreciated in either better methods of collecting within the loop or in transforming the cell array to a double matrix.

Answer 1

[Image Analyst]

Don't use "all" as the name of your variable - it's the name of a built in function. Not sure if that is causeing your problem or not but you should not use that name!

Let's say you want the matrix you stored in cell #42 of a cell array called "ca":

theMatrix = ca{42};

Now let's say you want the element at row 3, column 2:

theValue = theMatrix(3, 2);

This should explain it pretty well: http://matlab.wikia.com/wiki/FAQ#What_is_a_cell_array.3F

A structure array is simpler so maybe you'd prefer that over a cell array (which has very high overhead).

[Bus141]

I understand how to get the value from a specific cell but my problem has thousands of columns and creating a matrix for each cell would be excessive.

[Stephen23]

Note that you should preallocate the cell array, as expanding it on every iterations is very inefficient and can cause memory problems:

http://www.mathworks.com/help/matlab/matlab_prog/preallocate-memory-for-a-cell-array.html

https://www.mathworks.com/matlabcentral/newsreader/view_thread/292183

[Image Analyst]

I'm not the one who told you to put thousands of different sized matrices into cells - that was your decision. I merely told you how to access them after you stuck them in there. Cell arrays are flexible but very memory inefficient. If you think it's excessive, then don't do it. Use another variable type if you can. Can you use a structure array like I suggested? Or else allocate a 3D array with lateral dimensions that are the largest you ever expect you matrices to be. Do you know the max size they will ever be?

[Bus141]

The problem is that each iteration gives me a different sized matrix of numbers. I am new to structured arrays but will look further into it. The issue is that I want to extract all of the different sized matrices at the end of the loop and this number is larger than one would want to manually search through but not so large where pre-allocating for efficiency is a concern.

The 3-D dimensions idea gave me the thought to pre-allocate a larger matrix and then fill in what I get with each iteration.

for i=1:10
    within_z=zeros(10,1);
    z=randi(10,1);
    store=randi(200,z,1);
    l=length(store);
    within_z(1:l,:)=store';
    alt(:,i)=within_z';
 end

Thanks for the help.

[Image Analyst]

Making a 3D array could work, but you might want to have a "rows" array and a "columns" array to keep track of what part of the slice was actually used. Or else initialize unused elements to some flag value, like -99999 so that you can recognize what's been assigned and what's not been assigned at that slice/plane/z-level. This method would take up a LOT less memory and be much faster. Even though it "wastes" some memory, it's not as wasteful as a cell array with it's huge overhead. If it works for you, maybe you could Vote and/or Accept the answer.