Replace each element of a matrix with the number it is closest to in a set of discrete values without using for loop

I have a set of discrete values in the row vector <list>. I want to take a vector M of real doubles, and for each element M(j) in M I want to replace M(j) with the element in <list> that is closest to M(i) on the number line. For example, if my list was [2 4] and my matrix was [1.1, 3.2, 5, -17], the resulting matrix would be [2, 4, 4, 2].

I know I can do this with a for loop:

for j=1:length(M)
    [m mn]=min(abs(list-M(j)));
    M(j)=list(mn);
end

Is there any way to do this without a for loop?

Answer 1

[Andrei Bobrov]

A = [2 4], B = [1.1, 3.2, 5, -17]

[~, ii] = min(abs(bsxfun(@minus,B(:),A(:).')),[],2); % MATLAB <= R2016a
[~, ii] = min(abs(B(:) - A(:).'),[],2);              % MATLAB >= R2016b
out = A(ii)

or

F = griddedInterpolant(A,'nearest');
out = F(B)

or

out = interp1(A,B,'nearest','extrap');

[Ted]

Ah I did not know about this bsxfun, thank you!

Answer 2

[MG Poirot]

I think I've got a simple solution for you by finding the minimal difference after matrix subtraction:

list = [2 4]';
M = [1.1 3.2 5 -17];

A = repmat(M,numel(list),1)
B = repmat(list,1,numel(M))
dif  = abs(A-B)
[~, I] = min(dif,[],1)
list(I)'

-mgp