The problem is not in the concatention; it is in the display of the results. Apparently all the entries from A that are shown are smaller in magnitude than 0.00005 so that when rounded to four places you see a zero displayed. Try using "format long" before displaying the results. Or better still use floating point formatting provided by fprintf to allow each entry to be displayed in accordance with its own magnitude. Your first row might then look something like this:
1.111111111111111e-10 -2.222222222222222e-11 1.634987654321234e-1
Note: I am responding here to your original posting. As you have now edited it the results look impossible. You had better check that again.
