How to remove duplicate nodes and those inbetween from an array of elements?

I really need help on this. Been at it for a week.

I have a randomly generated path between nodes 1 to 25.

Initial generation: 1 6 7 2 1 6 7 8 9 4 3 2 1 6 7 12 11 16 21 22 23 24 19 20 25

I want to remove all repeated segments and nodes so based on this, the final path I want is:

(1, 6, 7, 12, 11, 16, 21, 22, 23, 24, 19, 20, 25)

How should I do this? Pls help.

A user helped me reduce this with the following code:

while true
 pathhist = histcounts(path, [unique(path), Inf]); 
 duplicate = path(find(pathhist > 1, 1)); %find identical digit
 if isempty(duplicate)
    break; %nothing left to remove, exit loop
 end
 occurences = path == duplicate;
 path(find(occurences, 1) : find(occurences, 1, 'last')-1) = []; %delete 1st occurence and anything up to last occurence
end

However, my paths generated may have up to 500 elements and the code doesnt seem to be able to handle it.

Help is really, really appreciated

[KSSV]

if you want to remove repeated nodes, why dont you use unique?

[Walter Roberson]

Over in http://www.mathworks.com/matlabcentral/answers/275204-how-to-remove-duplicate-nodes-from-generated-path-and-segments-inbetween, Guillaume is waiting for you to give an example of where it would run forever.

[Xi Wen Lim]

@Walter,

Thank you for the notice. I did not realise Guailaume had responded

Answer 1

[Walter Roberson]

Do not use "path" as a variable name, as that interferes with the MATLAB path.

Calling the generated path "gpath", then:

IDX = 1;
while IDX <= length(gpath)
  dpos = find(gpath==gpath(IDX),1,'last');
  if dpos > IDX
   gpath(IDX+1:dpos) = []; 
  end
  IDX = IDX + 1;
end

This is guaranteed to take no more than length(gpath) steps.

[Xi Wen Lim]

Thank you very much Walter! I have no problems with running this code even with many elements.