MYUNIQUE

finds uniqe entries, indices and number of occurances of the unique elements

Description

MYUNIQUE is a function to find unique elements in a numeric vector in the order in which they appear first or last which is determined by "mode" (default is first). The function is also outputs the indices "ndx" of unique elements and number of occurances of that elements in "h".

% Example:
% >> a=[9 3 6 2 7 1 2 1 1 1 9 9 3 3];
% >> [b ndx h]=myunique(a)
% b =
%
% 9 3 6 2 7 1
% ndx =
%
% 1 2 3 4 5 6
% h =
%
% 3 3 1 2 1 4
%
% >> myunique(a,'last')
% ans =
%
% 6 7 2 1 9 3

• MATLAB release: MATLAB 7.4 (R2007a)

[Richard Brown]

Yet another "myunique" posted to the file exchange, which is the same as unique, but with unsorted output. This is the third in a week - is this a homework assignment?

Stating that it returns "unsorted" output sounds to anyone who reads it that it should perform better than unique as it doesn't sort the values.

This is in fact not the case, what the code does is return the unique elements in the order in which they appear, which takes more effort than unique (two sorts used in this code, for instance). The code can also return the indices of these first (or last) appearing values but this isn't stated in the help, although this would be probably one of the submission's most useful features.

Clear up the documentation, and state the point to improve it

>> x = floor(100 * rand(1, 1000000));
>> tic, a = unique(x); toc
Elapsed time is 0.301613 seconds.
>> tic, b = myunique(x); toc
Elapsed time is 0.730075 seconds.

[JK L]

very useful

[José-Luis Guerrero]

Works like a charm

[Reza Farrahi Moghaddam]