how to find the sum of multiple non zero elements in an array

hi, i have an array that has like zero and non zero elements like eg: 0 0 0 1 1 1 0 0 0 3 3 3 0 0 0

so i want output i=3 and j=9 that is the sum of first set on nonzero elements seporate and the second seporate

[Jan]

Is "i" and "j" really required? Note that this might get extremely complicated, when you have 5000 such blocks and have to avoid collisions with other existing variables.

Answer 1

[Jan]

a      = [0 0 0 1 1 1 0 0 0 3 3 3 0 0 0];
c      = cumsum(a);
index  = strfind([a,0] ~= 0, [true, false]);
result = [c(index(1)), diff(c(index))];

Not tested yet!

[Mitson Monteiro]

thanks buddy, for ur answers.... what if i want the index values of the non zero numbers so that i can minus the first index from the last index of first set and the second set and if there are more than one colum data

[Jan]

@Mitson: Do you mean something like this:

index  = strfind([0, a] ~= 0, [false, true]) + 1;

Answer 2

[Roger Stafford]

There is an ambiguity in the posing of this problem. Are the sets of nonzero numbers distinguished from one another by their values or by their contiguity? For example, if your vector were [1 3 0 3 1] do you want [2 6] (distinguished by values) as a result or [4 4] (distinguished by contiguity)? The following assumes the latter.

Let x be the row vector of your numbers.

 t = [0,x,0];
 c = cumsum(t);
 f = find(diff(t~=0)~=0);
 r = c(f(2:2:end))-c(f(1:2:end));

My apologies to you, Jan, if this seems too similar to your solution. I couldn't resist giving it.

[Matt J]

For example, if your vector were [1 3 0 3 1] do you want [2 6] (distinguished by values) as a result or [4 4] (distinguished by contiguity)? The following assumes the latter.

Yes, and my solution assumes the latter.

[Roger Stafford]

Don't you mean the former, Matt?

[Matt J]

Um. Yes, the former.

Answer 3

[Matt J]

[u,i,j]=unique([0 0 0 1 1 1 0 0 0 3 3 3 0 0 0])

result = histc(j,1:max(j)).*u

Answer 4

[Mitson Monteiro]

thanks buddy, for ur answers.... what if i want the index values of the non zero numbers so that i can minus the first index from the last index of first set and the second set and if there are more than one colum data