Better way to calculate the SUM of one, two and three consecutive items

Question

class

public class myItem
{
    public int ID;
    public long Value;
    public long Sum1;
    public long Sum2;
    public long Sum3;
}

data

ID   Value
1    25
2    45
3    56
4    21

result:

     Sum1     Sum2         Sum3
1    25     25 + 45    25 + 45 + 56
2    45     45 + 56    45 + 56 + 21
3    56     56 + 21    56 + 21 + 0
4    21     21 + 0     21 +  0 + 0

Current procedure: Works but very slow (~10 minutes) with 100k rows.

List<myItem> list;
for (int i = 0; i < list.Count(); i++)
{
   myItem m = list[i];
   m.Sum1 = list.Where(x => x.ID == i).Sum(x => x.Value);
   m.Sum2 = list.Where(x => x.ID >= i && x.ID <= i + 2).Sum(x => x.Value);
   m.Sum3 = list.Where(x => x.ID >= i && x.ID <= i + 3).Sum(x => x.Value);
}

So I guess there should be a way to do it without the for to speed things up.

Answer 1

The root cause of the slowness is that your loop is O(n^2), since for each element you search the entire list for its successors. Below reduces that to O(n log n) (slowest part is the sort).

List<myItem> list;
list.Sort(<.. sort by id ..>);
for (int i = 0; i < list.Count; i++)
{
   myItem m = list[i];
   m.Sum1 = m.Value;

   m.Sum2 = m.Sum1;
   if (i < list.Count - 1)
   {
       m.Sum2 += list[i + 1].Value;
   }

   m.Sum3 = m.Sum2;
   if (i < list.Count - 2)
   {
       m.Sum3 = += list[i + 2].Value;
   }
}