how can i create a loop or function

I have an Excel sheet with many columns and rows.

Say column “y” has serval “groups” of the same numbers but the groups are any length, I want to find the length of each group of the same numbers then use that information to sum the value of a vector in column “k” The following code works, for the 1st three groups but there are 20+ groups in the column and I need to find a more efficient way to get the answer

[numbers, strings, raw] = xlsread('BCT2.csv');
colk=  numbers(:,11);
coly =numbers(:,25);
cf=coly;
EI=colk;
z=cf-cf(1); %%subtracts the first value from vector cf
idx = find(z<=0);
x=length(idx);
y=1:x;
A=sum(EI(y));
v=x+1;% the new starting point
z=cf-cf(v);
idx = find(z<=0); % indices
x1=length(idx);
d=x1-x;
ya=1:d;
B = sum(EI(ya))-A;
v2=x1+1;
z=cf-cf(v2);
idx = find(z<=0); % indices
x2=length(idx);
d1=x2-x1;
ya=1:d1;
C = sum(EI(ya))-B;

Answer 1

[Stephen23]

An easy way to detect groups of values is using diff and cumsum.

Then you could use accumarray, which is intended for exactly this kind of operation. Here is a simple example, where each run of identical digits in A is used to sum the values in B (e.g. the first run of A (the nines) sum the first B values [1,2,3] to give six:

>> A = [9,9,9,1,2,2,1,1]';
>> B = [1,2,3,4,5,6,7,8]';
>> X = cumsum([1;0~=diff(A)]);
>> accumarray(X,B)
ans =
     6
     4
    11
    15

If you would prefer to group by value rather then by runs of one values, then you can use unique:

>> [C,~,Y] = unique(A);
>> accumarray(Y,B)
ans =
    19
    11
     6
>> C
C =
     1
     2
     9

Note in this case the ones are summed together, even though they occur in two different runs in A.