Operation on a column with respect to other columns

Question

Looking at following code:

x=[2,3;1,4;1,5;2,4;3,1;3,4];
u=unique(x(:,1));
for i=1:length(u)
    n(i,:)=mean(x(x(:,1)==u(i),:))
end

Is there anyway to avoid the for loop, and make the code simpler? Basically, what I want to do is to group the rows based on a the value of a column and apply an operation on the other values for a given "group".

Possible duplicate of [MATLAB find and apply function to values of repeated indices](http://stackoverflow.com/questions/16086874/matlab-find-and-apply-function-to-values-of-repeated-indices) — marsei, Dec 22 '15 at 22:14

Benoit_11 · Accepted Answer · 2015-12-22T22:33:35.520

2

That's a job for the powerful accumarray with the mean function. Basically use values in the 1st column of x as subscripts and the values from the 2nd column as actual values on which you perform the desired calculation:

n2 = accumarray((x(:,1)),x(:,2),[],@mean)

out = [unique(x(:,1)) n2]

out =

    1.0000    4.5000
    2.0000    3.5000
    3.0000    2.5000

yay!

edited Dec 22 '15 at 22:33

answered Dec 22 '15 at 22:02

Benoit_11

13,905
2
24
35

1

Instead of writing `@(x) mean(x)` as function, it recognises `@mean` by itself, it being a built-in. No need of anonymising it. – Adriaan Dec 22 '15 at 22:05
1

oh right I keep forgetting about that thanks @Adriaan – Benoit_11 Dec 22 '15 at 22:33

Operation on a column with respect to other columns

1 Answers1