I have some data that looks like this:
+---+--------+-------------+---------------+--------------+
| | A | B | C | D |
+---+--------+-------------+---------------+--------------+
| 1 | row_id | disposal_id | excess_weight | total_weight |
| 2 | 1 | 1 | 0 | 30 |
| 3 | 2 | 1 | 10 | 30 |
| 4 | 3 | 1 | 0 | 30 |
| 5 | 4 | 2 | 5 | 50 |
| 6 | 5 | 2 | 0 | 50 |
| 7 | 6 | 2 | 15 | 50 |
| 8 | 7 | 2 | 5 | 50 |
| 9 | 8 | 2 | 5 | 50 |
+---+--------+-------------+---------------+--------------+
And I am transforming it to look like this:
+---+--------+-------------+---------------+--------------+
| | A | B | C | D |
+---+--------+-------------+---------------+--------------+
| 1 | row_id | disposal_id | excess_weight | total_weight |
| 2 | 1 | 1 | 0 | 30 |
| 3 | 2 | 1 | 10 | 30 |
| 4 | 3 | 1 | 0 | 20 |
| 5 | 4 | 2 | 5 | 50 |
| 6 | 5 | 2 | 0 | 45 |
| 7 | 6 | 2 | 15 | 45 |
| 8 | 7 | 2 | 5 | 30 |
| 9 | 8 | 2 | 5 | 25 |
+---+--------+-------------+---------------+--------------+
Basically, I need to update the total_weight column by subtracting the sum of the excess_weights from previous rows in the table which belong to the same disposal_id.
I'm currently using a cursor because it's faster then other solutions I've tried (cte, triangular join, cross apply). My cursor solution keeps a running total that is reset to zero for each new disposal_id, increments it by the excess weight, and performs updates when needed and runs in about 40 seconds. The other solutions I've tried took anywhere from 3-5 minutes and I'm wondering if there is a relatively performant way to do this using set based operations?