Collect OLS regression results

Hi guys, I am Ambrogio Dalò, an university student.I am using Matlab for my final dissertation and some commands are still not clear.

What I wont to do with the following script is : 1) Create a dependent variable matrix (SRW) 2) clear my dependt variable form colums with less than n oservation 3) regress the regressor matrix (X) on each dependete variable 4) collect all the results for each j in an cell-arrey matrix

I'll explain my script,in order to be as clear as possible.

for j = 1:245; % start the loop j

SRW = SR(j:j+23,:); % create the dependet variable matrix

A = sum(isfinite(SRW)); % sum the finite value for column

I = find(A<16); % find the colum with less than 16 observatios

SRW(:,I(1:end))=[]; % clear the SRW matrix

S = size(SRW); % find the size of SRW matrix

for i = 1:S(1,2); % start the second loop, whose dimension depend from the size of S

Y = SRW(:,i); % create a dependent variable for each column of SRW

X = horzcat(ones(24,1),Factors(j:j+23,5)); % create the regressors matrix

B = NaN(2,i); % preallocate hte Btas matrix

T = NaN(2,i); % preallocate the t-stat matrix

P = NaN(2,i); % preallocate the P-value matrix

R = NaN(1,i); % preallocate the R^2 vector

[B(:,i),T(:,i),P(:,i),R(:,i)] = NaNOLS(X,Y); % make the regression ignoring the NaN values

All = vertcat(B,T,P,R);

ToT(:,j)={All};

end end

Now, my problem is that the only results available are those for the last i-th regression for j=245.The ohters seem to disappear.I think becouse that for each regression matlab overwrite the results.In this way the only results available are the last one. Is there a way to solve this problem?

Thanks in advance for any help