MATLAB: Error using zeros “Size inputs must be integers”

zeros

Hi all,
I have a very bizarre problem with zeros. When typing:
zeros(10*(1-.7),1)
I get the following error:
Error using zeros
Size inputs must be integers.
Although using zeros(10*(.3),1) or zeros(10*(1-.6),1) doesn't cause any problem.

Best Answer

Floating point maths is not guaranteed to result in exact integer values. Just use the 'round' function if you need to use the result of such maths as an index into an array.
There are others with far greater expertise than me on this issue who may add comments/answers, but 0.7 may not be precisely represented meaning that 1 - 0.7 will not be precisely 0.3 (close enough for us not to care, but not precise enough to count as an integer when multiplied by 10).
Related Question