How to Sum Certain Values in Matrix A Based on Values of Matrix B in a Loop

Hello,

I have a 16 x 24 matrix A composed of one row of the hours 0,1,2...23 (the first row), and in the rest of the rows, I have hourly wind powers, in which each row corresponds to the wind speeds in a different location in the world.

What I am trying to do is use matrix B of size 15 x 2 to specify start and end times in matrix A in a loop, and then add up the wind powers in each row between those times, inclusive, and then store each of those sums as an entry in a 15 x 1 matrix. Each row of B has different start and end times.

I will illustrate with a example:

A = [0 1 2 3 4 5... 
     5 6 7 8 9 2...]

B = [0 3];

size(A) = 2 x 5;
size(B) = 1 x 2;

So I want to add up 5, 6, 7, and 8 to get 26, and store that value in a new column matrix. I want to do this for each row of A and B, but for different start and end values, depending on the row of B.

What I've done so far is this:

%%G is equivalent to matrix A above; Rounded_Bounds_WS is equivalent to matrix B above; J is the column matrix I'd like to create.

for s=1:length(files) %files is just the collection of .nc files I'm using (15)

      if and(G(1,:)>= Rounded_Bounds_WS(s,1), G(1,:)<= Rounded_Bounds_WS(s,2))

          J(s,:)=sum(G(s+1,:));

      end

end

But I haven't been able to get this to work. s cycles from 1 to 15 (where 15 =length(files); however the matrix J isn't even created.

Thank you in advance for your help.

Andrew

[edited for clarity -- the cyclist]

[Azzi Abdelmalek]

In your example you gave A and B, then you are referring to C and D, this is not clear

[Andrew]

I've edited my question to make that clearer.

Answer 1

[kjetil87]

Im not a 100% sure if this is what you asked for but ill have a go at it =)

A=[0:23;...  %hours 0 to 23
   rand(15,24)]; %15x24 submatrix with random values.

B=randi([0,23],15,2)]; % 15x2 matrix with numbers between 0 and 23
B=sort(B,2);           % now the first value in each row is the lowest.

Im not sure what "files" and length(files) is, but ill make an example based on your first A and B.

J=zeros(size(B,1)); %pre allocate J to avoid having to reallocate during loop.

for i=1:size(B,1)

   J(:,i)=sum( A(2:end,(B(i,1):B(i,2))+1),2);  % sum each row in a from B(i,1) 
                                               % to B(i,2).
                                               % add one to the index since 
                                               % "hour 0" in A
                                               % is located in column 0.

end

Is this what you wanted? If e.g B was equal to

B=[0,6;
   ..
   ..
   ..

J(1,1) will contain sum( A (2, 1:7) ) . While J(2,1) contains sum( A (3,1:7) ) and so forth.

To realete to your first example:

J=[];
A = [0 1 2 3 4 5; 
     5 6 7 8 9 2];
B = [0 3];

J(:,1)=sum( A(2:end , (B(1,1):B(1,2))+1 ,2)

J =

      26

Note that the sum is 26, because i assume that when B is [0,3] you want to use colums of A that contains 0 , 1 ,2 AND 3. If you only want 0 1 2 remove

   (B(i,1):B(i,2))+1

and replace it with

    B(i,1)+1:B(i,2)

In your specific example:

sum( A(2:end , B(1,1)+1:B(1,2) ) ,2)

ans =

      18

[Andrew]

Thank you so much! This seems to be exactly what I need- I will let you know if that isn't the case.

I actually mistyped before: for the range [0,3], I do want to add up the value that corresponds to 0, 1, 2 and 3. Before, I had written only that I wanted to sum the values corresponding at 0,1,2. I apologize.

Also files is just the collection of .nc files I am reading into my loop, and length(files) is 15, the number of files that are being read in.