Saving/converting cell arrays into a CVS file

I have an m*n matrix X of numerical data and want to save it to a cvs file with a 3x3 header format (i.e. 3 header rows, 3 header columns, then the data).

The matrix is big (talking >6000x6000), and the various headers are made up of the same few strings repeated in various orders.

I've started by making a cell array of the various header components,

eg header_comps={'string 1', 'string 2', 'string 3' ...}

and then from this can generate either one big cell array (header_all=cell(m+3,n+3)) with all the headers in and an empty space where X will go, or a number of individual cell arrays (row_header_1=cell(1,n), col_header_1=cell(m,1) etc.) to contain all of the required header info.

I can then either put X into the large (m+3,n+3) cell array, or have X and the various header cell arrays separate that when put together make up the desired result.

Where I am stuck is with how best to handle this data and write it to a csv file.

I am aware with a simple data matrix and row header you can use something like:

X=rand(3); %data matrix; 
header=['header obj 1, header obj 2, header obj 3']; 
outid = fopen('filename.csv', 'w+'); 
fprintf(outid, '%s', header); 
fclose(outid); 
dlmwrite ('filename.csv',X,'roffset',1,'-append')

which writes the char array row header into the file, then offsets the data X by one row.

I'd be happy to write 3 row headers in, 3 column headers in, then use coffset and roffset of 3. The problem is I don't seem to be able to generate the char header arrays in an automated way which I can with cell arrays, but I don't seem to be able to write cell arrays in in the same way you can a char array, and converting the cell arrays to char arrays seems to mess it up.

In short I have a list of strings that are used to generate the row and column headers. I'd like to generate row and column headers, and then put them and the data matrix in a csv file.

I've looked at a lot of solutions to writing cells to csv, but the best I've seen yet seems to be a cell by cell for-loop solution. There is a surely a more elegant way to do this?

Any help much appreciated. Regards, Simon