## The most efficient way to locate the first 1's in each row of a matrix

### Ulrik William Nash (view profile)

on 19 Oct 2019
Latest activity Commented on by Jos (10584)

### Jos (10584) (view profile)

on 21 Oct 2019
I have a potentially large r by c matrix, m, consisting of 1's and 0's. What is the most efficient way create a r by 1 vector, v, containing the index of the first 1 in each row of m.
For example,
m = [0 1 1;1 0 0; 0 0 1]
should result in
v = [2;1;3]

### Star Strider (view profile)

on 19 Oct 2019

The max function will return the indices of the first instances of the maximum value it encounters, in the desired dimension. Since your matrix only has (0,1), it is appropriate here.
This:
m = [0 1 1;1 0 0; 0 0 1];
[~,v] = max(m,[],2)
produces:
v =
2
1
3

Jos (10584)

### Jos (10584) (view profile)

on 21 Oct 2019
nice! One can shorten it to:
[~, v] = max(m')
... but what if there is no 1 in a row

### Jos (10584) (view profile)

on 19 Oct 2019

m = [0 1 1 1 ;1 0 1 0; 0 0 1 1 ; 0 0 0 1]
% a simple solution
[r, c] = find(m') ;
clear v1
v1(r,1) = c
% another solution
[v2, ~] = find(cumsum(cumsum(m,2),2)==1)