Cumsum at different starting indices, append all previous value.

Hello, say I have 2 continuous dataset for example:

A = [0 0 8 0]   B = [1 0 0 0 0.5 1.5 0]

The algorithm iterates over each element i = i +1 to check if element >= 0.5. If True cumsum starts.

Condition: Cumsum ends when exceeds 2 within 3 indices. Append the values and index.

The final cumsum value is not important, it is a means to set the condition of appending relevant values.

So result for A will give:

result = [0, 0, 8], idx = 1, 2, 3 NOT result = [ 0 0 8 '0']

However, if the condition is not met. Ignore the cumsum and move on to the next value.

e.g.: B will start 1st cumsum at [1]: 1 + 0 + 0 = 1. Which fails to meet the condition. Hence ignored.

It continues iterating, until reaching 0.5 which then starts another cumsum.

So result for B will give:

val = [1 0 0 0 0.5 1.5] NOT [1 0 0] [1 0 0 0 0.5 1.5 '0']

How can I approach this? Many thanks.

[Jan]

I want to append all values until an element >= 0.5, then cumsum happens.

A = [0 0 8 0]  
val = [0, 0, 8], idx = [1, 2, 3]

Here is no cumulative sum.

B = [1 0 0 0 0.5 1.5 0]
val = [1 0 0 0 0.5 1.5]

There are no added elements here also. If [1,0,0] is ignored, what happens with the 3rd zero?

Currently it is not getting clear to me, what you want to achieve.

[Jonathan Cheong]

Sure let me try to elaborate,

Say we iterate over each element: i = i + 1 to check if the element is >= 0.5

For A: Because both zero [0] [0] at index 1 and 2 is less than 0.5 the cumsum doesnt happen.

When it reaches element [8], which is greater than 0.5, it starts a cumsum.

8 + end because the cumsum >= 2, within 3 indices.

It appends all the previous value and the cumsum final value to storage.

Result = [0 0 8] NOT [0 0 8 '0']

For B: 1st element [1] is greater than 0.5 hence cumsum starts:

1 + 0 + 0 = 1. Because the sum does not exceed 2 and is within 3 indices, this cumsum is ignored.

Then the algorithm continutes to iterate the data, and starts cumsum at [0.5] because >= 0.5

[0.5] + [1.5] + end. Because the cumsum >= 2 within 3 indices.

Then, it appends all previous value until final cumsum value to storage. It doesn't include the final 0.

Result = [1 0 0 0 0.5 1.5] NOT [1 0 0] or [1 0 0 0 0.5 1.5 '0']

Note: The final cumsum value doesn't matter. It is just a means to set a condition so that correct data is appended.

Many thanks in advance

[Jonathan Cheong]

Here is my code, it doesn't work but hopefully this will help.

a = 0;
dd = [];
ind = [];
for row = idx
    a = a + 1;  % Iterate over each indices     
    if rain(idx(a)) >= 0.5   % If value >= 0.5 start cumsum
        cs = cumsum(rain(idx(a):end);
        cmp = find(cs <= 2); %  Find index where cumsum ends
        if numel(cmp) > 3
            continue         % If index >3, ignore this interval            
        end
%  Append all relevant vlaues?
        ind2 = a + cmp;
        ind = [ind; ind2];
    end
end

The files are also attached below.