Combining data with matching elements in the first two columns

Hello, I have a data table that looks like this

920  381  784  0
920  381  0  21.4375
23  388  1703  0
23  388  0  4.109375
445  487  304  0
445  487  0  15.09375
1100  506  1480  0
1100  506  0  28.234375
245  520  454  0
245  520  0  40.21875

For all the entries where the first two columns match (ie, the first two rows), I would like to combine those two rows into one. So these two rows:

920  381  784  0
920  381  0  21.4375

Will become

920  381  784  21.4375

And so on for the rest of the data set. I would appreciate any help.

Thanks

[the cyclist]

Is that exact pattern guaranteed? Namely, will the first zero always be in column 4, and the second zero always be in column 3? And will there always be a pair of rows like that?

[Muneer]

No it isn't guaranteed that there will always be a pair of rows. If there is one row with no matching row, however, it will have a value in column three, not column four. Thanks

Answer 1

[the cyclist]

Here's one way:

[~,i,j]=unique(M(:,1:2),'rows');
[M(i,[1 2]),accumarray(j,M(:,3)),accumarray(j,M(:,4))]

where M is your original array.

EDIT: Original solution only gave the last two columns, so I fixed it to give all the columns you need.

[Muneer]

Thanks for the help!

[Azzi Abdelmalek]

This does not give the result

[Azzi Abdelmalek]

The cyclist, your code doesn't work if there are more then two duplicate rows (I mean duplicate for the two first column).

Answer 2

[Azzi Abdelmalek]

M=[920  381  784  0
  920  381  0  21.4375
  23  388  1703  0
  23  388  0  4.109375
  445  487  304  0
  445  487  0  15.09375
  1100  506  1480  0
  1100  506  0  28.234375
  245  520  454  0
  245  520  0  40.21875]

[~,idx,jj]=unique(M(:,1:2),'rows','stable'); % M is your matrix
n=numel(idx)
out=zeros(n,size(M,2))
for k=1:n
  out(k,:)=max(M(find(jj==k),:))
end

[Muneer]

I don't think this worked. It returned a matrix with the right number of rows, but the fourth column was filled with values of 0 only, not the data I needed combined in.

[Muneer]

I think it might have just looked at the first two columns, taken the first result if there was a duplicate and deleted the second row without combining the data in.

[Azzi Abdelmalek]

Look at edited answer

[Muneer]

Thanks for the help!