Use of 2012 unique version with Matlab 2011a possible?

Would it be possible to use the new version of "unique" in an older matlab version? I wrote some code and used "unique" with the 'stable' option and I need to run this code on another computer with Matlab 2011a.

If it's possible: Which files would I need to copy?

[Jan]

See also: http://www.mathworks.com/matlabcentral/answers/30956-change-of-set-function-in-the-future

Answer 1

[Walter Roberson]

This would not be permitted under the license agreement, and would not be technically supported.

Answer 2

[Jan]

function [AA, AI, BI] = CStrUnique(A)
% Unique elements of a cell string [MEX]
% [AA, AI, BI] = CStrUnique(A)
% INPUT:
%   A:  Cell string of any size.
% OUTPUT:
%   AA: Cell of unique strings of A. If the input A is a column cell, AA is a
%       column cell also. Otherwise AA is a row cell. The comparison considers
%       spaces, empty strings and case of letters.
%   AI: Linear indices (see IND2SUB) of unique strings such that A(AI) is
%       unique. The first occurrence of a string is preferred.
%       The order of AI corresponds with the order of strings in A, so AI is
%       strict monotonically increasing.
%   BI: List of linear indices of input strings according to output:
%       AA(BI) = A.
%
% Differences to Matlab's UNIQUE:
%   Faster.
%   Output is not sorted.
%
% Tested: Matlab 6.5, 7.7, 7.8, 7.13, WinXP/32, Win7/64
% Author: Jan Simon, Heidelberg, (C) 2006-2012 matlab.THISYEAR@nMINUSsimonDOTde
% License: BSD, use, copy, modify on own risk, mention the author

nA = numel(A);
if nA > 1
   [As, SV] = sort(A(:));

     if nargout < 3
        UV(SV)  = [1; strcmp(As(2:nA), As(1:nA - 1)) == 0];
        AI      = find(UV);
     else  % Indices requested:
        UV      = [1; strcmp(As(2:nA), As(1:nA - 1)) == 0];
        UVs(SV) = UV;
        AI      = find(UVs);

        % Get BI such that AA(BI) == A:
        v      = zeros(1, nA);
        v(AI)  = 1:length(AI);    % Sequence related to AA
        vs     = v(SV);           % Sorted like A
        vf     = vs(find(vs));    %#ok<FNDSB> % Just the filled entries
        BI(SV) = vf(cumsum(UV));  % Inflate multiple elements
     end
  elseif nA  % Comparison of subsequent elements fails for nA == 1
     AI = 1;
     BI = 1;
  else
     AI = [];
     BI = [];
  end

AA = A(AI);

Example: If the input is the (already unique) list of 961 M-files with full path, this is 18% faster than unique - and unsorted.

[Christopher]

Thank you very much Jan Simon.

[Jan]

Perhaps, I'm not shy enough, I know, it would be worth to vote and/or accept this answer. Walter's answer is more correct, but my one is more helpful.

Answer 3

[Christopher]

Thank you for your fast answer. Then I have to adjust my algorithm in order to fix this problem.

[Geoff]

That's probably the easiest. Unless you don't rely on the subtleties. For instance, if you don't care which index you get for each unique value you can simply get the stable ordering back..

%%

[~, I] = unique(data);

U_stable = data(sort(I));

%%

Otherwise, you could always write your own slower version of unique in MatLab and use that if the version is less than 2012a, or indeed write one in C with good performance.

[Christopher]

Thank you very much Geoff.