Replace vector duplicates with unique numbers

I have a D-dimension vector and I want to obtain a vector where duplicates are replaced by random numbers in [1,D]. The final vector must have all different values from 1 to D preserving non-duplicates order.

For example take D=6:

[2,4,5,4,1,2]

Isolate the unique numbers:

[2,4,5,-,1,-]

Then put the remaining 3 and 6 values:

[2,4,5,6,1,3] or [2,4,5,3,1,6]

What's an efficient way to do it?

Answer 1

[James Tursa]

One way:

D = [2,4,5,4,1,2]
n = numel(D)
C = setxor(D,1:n)
[U id] = unique(D)
ix = setxor(id,1:n)
D(ix) = C(randperm(numel(ix)))

D =
     2     4     5     4     1     2
n =
   6
C =
     3     6
U =
     1     2     4     5
id =
     5
     1
     2
     3
ix =
     4
     6
D =
   2     4     5     6     1     3

[Giovanni Rosso]

Thanks :)