Merging two vectors (via horzcat) turns everything into zero

Hey everybody! I've stumbled across a serious problem and I cannot fathom if I'm just being too stupid. I have 19 variables called QM1, QM2, QM3 etc., all having one row but different number of columns, such as 1x40 or 1x50. They're all of double-type and consist of values ranging from -1x10^1 to -1x10^8, so all non-zero numbers. I'm trying to merge all of them into one big vector, plotting their absolute value over the corresponding velocities that I have as well on a double-log scale and creating a line of best fit. However, when either using

QMall=horzcat(QM1,QM2,QM3,...,QM19);

or

QMall=[QM1 QM2 QM3 ...];

and then looking at the values of the newly formed vector (which has at least the correct dimension of 1x718), suddenly 99% of the numbers have turned to (1.0e+09 *) 0.0000. The manual creation of the big vector seems to work if I do it gradually, such as QMall=[QM1,QM2,QM3] and then QMall=[QMall QM4 QM5] and so on up until the last two or three variables which then turn into zeros. Has anyone ever encountered this problem before? What I don't get is that it used to work in the beginning, but suddenly it doesn't. Your help would be much appreciated.

Answer 1

[Orion]

Your values didn't became 0.

it's just the display which is not complete.

if you do

a=1e-9;
b=1000;
c= [a b];
disp(c)
   1.0e+03 *

      0.0000    1.0000

a is still 1e-9.

Change the display format

format long
disp(c)
   1.0e+03 *

     0.000000000001000   1.000000000000000

it's just visual, the value in unchanged.

[Ben]

Well, that's embarassing. I did think of that of course, but when I tried to plot the data, it looked as if almost all the y-data was zero, too. Only now I noticed that the y-scale was by far too large to show the details. But thank you.