Update matrix using vector of column index

I have a 2-dim matrix like this:

    0.3436    0.3283    0.3890    0.0274
    0.1137    0.4443    0.3484    0.0889
    0.2748    0.3124    0.4074    0.0439

and a vector with column indexes for each row

     3
     2
     2

Task is to invert values to get result like this:

    0.3436    0.3283   -0.3890    0.0274
    0.1137   -0.4443    0.3484    0.0889
    0.2748   -0.3124    0.4074    0.0439

I believe there is a simple way to do that efficiently without for-loop cycles. Any idea?

Answer 1

[the cyclist]

Looks like a job for linear indexing:

M = [0.3436    0.3283    0.3890    0.0274
     0.1137    0.4443    0.3484    0.0889
     0.2748    0.3124    0.4074    0.0439];

v = [3 2 2];

[m,n] = size(M);

linearIndex = sub2ind([m n],1:m,v);

M(linearIndex) = -M(linearIndex)

Note that the 2nd and 3rd arguments to sub2ind have to be the same shape, so make sure they are both column vectors, or both row vectors, but not mixed.

[Georgio Mazarin]

Great thanks! It works!