Finding Duplicate Values per Column

Greetings, suppose Column A has these values - 7 18 27 42 65 49 54 65 78 82 87 98

Is there a way to compare the values (row by row) and search for duplicates? (I'm using Matlab R2010b)I don't want the duplicated values to be removed.

Thanks.

Answer 1

[Jan]

A = [7 18 27 42 65 49 54 65 78 82 87 98];
[n, bin] = histc(A, unique(A));
multiple = find(n > 1);
index    = find(ismember(bin, multiple));

Now the values A(index) appear mutliple times.

[Teik Yee]

thx

[Faheem Ud Din]

thank u sir G

Answer 2

[the cyclist]

Here's a slightly different way:

X = [1 2 3 4 5 5 5 1];

uniqueX = unique(X);
countOfX = hist(X,uniqueX);
indexToRepeatedValue = (countOfX~=1);

repeatedValues = uniqueX(indexToRepeatedValue)
numberOfAppearancesOfRepeatedValues = countOfX(indexToRepeatedValue)

[Harold]

Is there a way that I can use this approach for a cell array. Suppose I have a 1x4 cell array and each cell contains a (x,y) co-ordinates. Can I somehow modify your solution to search for duplicate cells and get rid of them?

[Yowh]

Harold can you give an example of your cell array variable ?

[Jan]

@Yowh: It is unlikely, that Harold listens to comments after 8 months.

[Anurag Pujari]

Accurate. What an excellent piece of code.

Answer 3

[Hannes Greim]

you can use "tabulate" for cell arrays.

[Clemens]

Working only for versions after 2012 with the intro of tables.

Answer 4

[Wesley Allen]

Duplicate Finding with Tolerance

If you want to find duplicates with tolerances (e.g., for non-integers), I use the following:

A = [1.313;2.4;2.400000001;1.31299999999;2.25;2.25;2.25000000001;3.7];
TOL = 1e-5;

uniqueA = uniquetol(A,TOL);
duplicateBool = abs(repmat(A,size(uniqueA.'))-repmat(uniqueA.',size(A))) < max(abs(uniqueA))*TOL;
duplicateCount = sum(duplicateBool).';

Just like with the cyclist's answer, if you want to isolate only the values that have more than one instance:

iDuplicate = (duplicateCount ~= 1);
repeatedValues = uniqueA(iDuplicate);
numberOfAppearancesOfRepeatedValues = duplicateCount(iDuplicate);

repeatedBool = duplicateBool(:,iDuplicate);

Using the Results

The unique values are in uniqueA:

>> uniqueA

uniqueA =

      1.3130
      2.2500
      2.4000
      3.7000

The quantity of each unique value is in duplicateCount:

>> duplicateCount

duplicateCount =

       2
       3
       2
       1

To get the indices of A corresponding to the n-th unique value, uniqueA(n)

>> n = 2;
>> uniqueA(n)

ans =

      2.2500

>> duplicateIndex = find(duplicateBool(:,n))

duplicateIndex =

       5
       6
       7

Answer 5

[Fernando Meo]

Here is another answer (a one liner)

If AA is a 2D matrix and you wish to find the rows which have a duplicate values in its columns,

RowsWhichHaveDuplicates = find(arrayfun(@(i (~isequal(length(unique(AA(i,:))),size(AA,2))), [1:size(AA,1)]));

Example

AA = [6   7  11  6; 7  11  4  8; 11  15  1  10; 15  4  14  12; 
18  13  18  8; 12  13  18  1; 3  14  6  18];

>> RowsWhichHaveDuplicates = RowsWhichHaveDuplicates = find(arrayfun(@(i) (~isequal(length(unique(AA(i,:))),size(AA,2))), [1:size(AA,1)]))

RowsWhichHaveDuplicates =
  1   5

If your values are real, then a tolerance can be set by using the matlab "round" function to the decimal places you wish to use.

AA = round(rand(10)*10,1); % First decimal place

AA =
    6.0000    2.0000    0.4000    6.7000    9.4000    0.6000    8.3000    3.1000    1.0000    3.0000
    9.1000    7.5000    0.6000    6.0000    0.7000    3.1000    0.3000    4.2000    9.0000    3.7000
    2.5000    8.9000    5.0000    3.4000    7.2000    6.6000    8.4000    9.3000    9.0000    7.6000
    8.6000    1.0000    4.1000    4.0000    8.3000    4.6000    2.6000    0.6000    0.8000    3.1000
    7.6000    5.2000    2.2000    3.9000    7.3000    0.2000    6.6000    8.2000    5.2000    9.6000
    2.2000    6.0000    4.3000    7.0000    5.1000    6.9000    6.7000    6.4000    2.8000    2.1000
    4.2000    9.8000    9.5000    1.4000    5.2000    4.1000    2.6000    8.2000    8.8000    7.3000
    1.3000    6.7000    2.0000    3.8000    7.6000    5.7000    3.3000    3.3000    6.7000    2.5000
    9.2000    8.5000    7.1000    2.2000    6.3000    9.9000    2.5000    9.5000    1.2000    8.9000
    2.9000    1.7000    7.8000    4.1000    0.7000    8.6000    7.1000    9.1000    3.7000    7.1000

RowsWhichHaveDuplicates = find(arrayfun(@(i) (~isequal(length(unique(AA(i,:))),size(AA,2))), [1:size(AA,1)]))
RowsWhichHaveDuplicates =
     5     8    10

Hope this helps