MATLAB: How to calculate the mean of certain rows of a matrix based on specific column values

logical indexing index column means residual subtract

I have a matrix with 2 columns: example: A=[ 123 1; 444 2; 634 1; 311 1; 111 2; . . . 222 1; 312 1;]
First.I want to find the mean of the values in column 1 based on the values of column 2. Means for value 1(mean1) and for value 2(mean2).
Second. Ι want to subtract the means from the respective rows.The mean1 from the rows in column 1 that correspond to value 1 in column 2 and the mean2 from the rows in column 1 that correspond to 2 value in column 2. And take a column of the residuals without spoiling the sequence of the values.
thanks!

Best Answer

This works:
A = [ 123 1; 444 2; 634 1; 311 1; 111 2; 222 1; 312 1];
Amean = accumarray(A(:,2), A(:,1), [], @mean)           % Calculate Means
AmeanMtx = [Amean [1;2]]                                % Amean For The Appropriate Column #2 (For Demonstration Only, Can Be Deleted)
Aresd = A(:,1)-Amean(A(:,2))                            % Use The Second Column To Index ‘Amean’ To Calculate The Residuals
Amean =
        320.4
        277.5
AmeanMtx =
        320.4            1
        277.5            2
Aresd =
       -197.4
        166.5
        313.6
         -9.4
       -166.5
        -98.4
         -8.4
EDIT — Added comment documentation.
Related Question