1-d is sufficient to explain here.
x = [1 2 3 3 4 5];
y = [1 2 3 4 5 6];
There is a replicated x, with two distinct values of y. Exactly what do you think "retaining the duplicate" means here? What value of y should be interpolated when x==3?
A property of interpolation is that for any given x, if one of the original data points is interpolated, you should get out exactly the same value as went in. (At least, exact to within floating point trash.)