Most efficient way to generate an array around a central value

Hi!

I am trying to span a linearly spaced array around a center value without knowing number of points in the array beforehand. I only have the step size and the lower and upper limits that must not be exceeded. For example, if center = -4,3, stepsize = 2, lower = -7, upper = 3, then the resulting array should be array = [-6.3 -4.3 -2.3 -0.3 1.7].

Is there a way to generate such an array as efficiently as possible (e.g. without while-loops)?

Thanks for all the help in advance!

Answer 1

[Jan]

A bold guess:

center = -4.3
step   = 2
low    = -7
high   = 3

vhigh  = center:step:high;
vlow   = center - step:-step:low;
result = [vlow(end:-1:1), vhigh]

Or in one step:

L = low  + rem(center - low,  step);
H = high - rem(high - center, step);
result = L:step:H

I'm not sure if this is more stable considering rounding errors:

result = linspace(L, H, 1 + (H-L) / step)

Answer 2

[Adam]

[ flip( center:-stepsize:lower ), ( center + stepsize ):stepsize:upper ];

looks like it would work, though I haven't tested all edge cases and there are probably neater ways.

If you prefer you could just do the simpler:

unique( [ flip( center:-stepsize:lower ), center:stepsize:upper ] );

to get rid of the duplicated centre point, though I imagine it is less efficient.