Finding non-unique values in an array.

I have an array that I would like to extract the non-uniques rows from...

'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [1]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [3]   [25]

So I would like to find the rows that have non-uniques values in the fifth (last) column so I get...

'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [1]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [0]

Thanks in advance!

[Sean de Wolski]

This should be a model for a well asked question +1!

Answer 1

[Sean de Wolski]

C = {'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [1]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [3]   [25]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [0]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [4]
'\\campus\home\home2014...'    's-pl T2* FGRE'   [1]   [2]   [4]
};

% Unique values
[~,idxu,idxc] = unique(cell2mat(C(:,5)));

% count unique values (use histc in <=R2014b)
[count, ~, idxcount] = histcounts(idxc,numel(idxu));

% Where is greater than one occurence
idxkeep = count(idxcount)>1;

% Extract from C
C(idxkeep,:)

[Shaun]

Thanks for your response. Using the code above but with histc in place of histcounts for my version of Matlab I get an error code:

Error using histc
Too many output arguments.

Any ideas?

[Sean de Wolski]

The histc equivalent would be:

[count,idxcount] = histc(idxc,1:numel(idxu))

[dpb]

I don't have histcounts either so and didn't go look it up...the needed indices using the method w/ histc above are

idxkeep= find(bin==find(n>1));

Answer 2

[dpb]

If the column of interest is y, then

u=unique(y);         % the unique values
[n,bin]=histc(y,u);  % count how many of each and where
ix1=find(n>1);       % index to bin w/ more than one

Index into y is bin(ix1) for each element in ix1

idx=[];
for v=find(n>1).'
  idx=[idx;find(bin==v);
end

Trial data...

>> y=[0 0 25 3 3 1].';
>> u=unique(y);
>> [n,bin]=histc(y,u);
>> for v=find(n>1).',idx=find(bin==v),end
idx =
   1
   2
idx =
   4
   5
>>

Looks correct when accumulate into the array...an accumarray expression ought to be workable but will leave as "exercise for the student"... :)