Inefficient code - simple counter

I have the following inefficient bit of code.

for i = 1:n %arb n
    if something
        A(i,1) = A(i,1) + 1;
    elseif somethingelse
        A(i,2) = A(i,2) + 1;
    end
end

There are actually 5 if/else clauses inside a function which is called inside the loop, but I've reduced it to the above for simplicity. For large n, this is very slow. Is there a faster way to do this? It's supposed to be a simple counter that retains information about the counts.

Thanks

edit: I have filled A with zeros before the loop also.

[KL]

What are your conditions? Give an example.

[Joppy]

I'm checking if one number is smaller than another. While they do contribute to the overall time, it isn't nearly as problematic as the counter (I've timed with and without).

Answer 1

[KL]

I suppose your condition is something like checking the range of the specific element. Depending on what range they are in you want to perform something on that specific element.

A = rand(50,1);
ind1 = A<0.25; %first if
ind2 = A>=0.25&A<0.5; %elseif
ind3 = A>=0.5&A<0.75;
ind4 = ~ind1&~ind2&~ind3; %else 
B = (A+1).*ind1+(A+2).*ind2+(A+3).*ind3+(A+4).*ind4;

[Joppy]

Oh. No the conditions aren't related to A, but I can adapt this approach anyway.. Silly me. Thanks!

[KL]

Alright.

More on this link: https://www.mathworks.com/help/matlab/matlab_prog/vectorization.html

Answer 2

[Birdman]

One approach:

Consider a B vector with dummy data.

B=randi([1 10],20,1);A=zeros(numel(B),2);
ind1=find(B<5);%condition1
ind2=find(B>=5);%condition2
A(ind1,1)=1;A(ind2,2)=1;