Is it a precision mistake? Creating a vector goes wrong.

I am getting a very strange bug namely:

    bboxMxY = max(UVW(curPoints(logic),2));
    bboxMxZ = max(UVW(curPoints(logic),3));
    bboxMnY = min(UVW(curPoints(logic),2));
    bboxMnZ = min(UVW(curPoints(logic),3));
    Ycor = bboxMnY:bboxMxY;
    Zcor = bboxMnZ:bboxMxZ

There is nothing fancy taking place here. Just 2 mins and 2 maxes. The mins are -2.000 and -3.000. And the maxes are 2 and 3.000. So unfortunately when I do

Ycor = bboxMnY:bboxMxY

I get the equivalent of -2:1 and not -2:2. Why is this happening? I tried to use ceil,fix or add an eps but they do not work. What is really the reason behind this, and how can I bypass it?

[Adam]

You haven't included the code where you tried ceil or fix which are the obvious solutions (or round or floor)

Answer 1

[John D'Errico]

Your numbers are not true integers.

format short
x = [1.999999 2.0000001 2.99999 3.000000001]
x =
    2.0000    2.0000    3.0000    3.0000

See that even though it looks like x is the vector [2 2 3 3] we know it is not. In your case, you only think your data ranges between integer limits, it does not.

[Ioannis Nemparis]

Still this does not explain why round(min) or fix(min) would not work

[Guillaume]

fix rounds towards 0, so fix(-1.9999) will return -1, not -2.

round should work though and so would:

Ycor = floor(bboxMnY):ceil(bboxMxY);

If it does not, then attach sample data that reproduces the problem.