How to find duplicate values in 2d matrix?

Hello,

I have the following matrix

M =[    1540	1685	6.90000000000000	24.2000000000000	14.6000000000000;
        1540	1700	9.90000000000000	15.5000000000000	17.9000000000000;
        1540	1701	9.90000000000000	12.4000000000000	17.9000000000000;
        1540	1685	6.90000000000000	23.5000000000000	14.600000000000
        1540	1700	9.90000000000000	15	                17.9000000000000;
        1540	1701	9.90000000000000	13.3000000000000	17.9000000000000;
        1540	1680	6.90000000000000	26.2000000000000	14.4000000000000;
        1540	1689	6.90000000000000	26.1000000000000	14.6000000000000;
        1540	1680	6.90000000000000	24	                14.6000000000000;
        1540	1680	6.90000000000000	24	                14.5000000000000;
        1540	1688	7	                23.5000000000000	14.6000000000000;
    ]

And I would like to identify the duplicates of that matrix based on the first two colums and extract them to furher operations. Thus, I would like to have a way to get a sub-matrices of the form

M11 = [ 1540	1700	9.90000000000000	15.5000000000000	17.9000000000000;
        1540	1700	9.90000000000000	15	                17.9000000000000;
      ]
M22 =[ 1540	1680	6.90000000000000	26.2000000000000	14.4000000000000;
       1540	1680	6.90000000000000	24	                14.5000000000000;
     ]

How can I achieve this? Could I use findgroups for this purpose?

Thanks in advance

[KSSV]

Read about unique

[Jan]

This is a 2D matrix, not 4D.

Answer 1

[Rik]

You can indeed use findgroups. You should supply the first two columns as separate inputs.

G=findgroups(M(:,1),M(:,2));

Then I would suggest looping through the group IDs and fill a cell array. Don't use numbered variables, because they are hard to use in further analysis.

[Javier]

Thanks Rik,

To be sure that I understand the output of findgroups, G variable contains group identifiers right? Meaning that for instance all elements with same M(:,1) and M(:,2) share same identifier, is that correct?

Then I could do something like

for k=1:max(G)
    idx = G == k;
    subM{k} = [M(idx,1) M(idx,2) M(idx,3) M(idx,4) M(idx,5)]
end

and I could check for the differences?

[Rik]

That should work. You can also use ismember to generate the idx vector (in which case that would become a logical vector). You can also more generally use M(idx,:) instead of specifying each column separately.

Also, starting off with subM=cell(1,max(G)); is probably a good idea.