Since I am quite lost in this calculation What I have now is a huge dataset with figures like this:
| Item Number | 2019w01 | 2019w02 | 2019w03 | 2019w04 | 2019w05 | Average Sale per Week |
| 4030-450-56 | 0 | 0 | 2 | 0 | 3 | 1.6667 |
| 4030-451-55 | 5 | 6 | 7 | 0 | 0 | 3.6 |
| 3025-654-28 | 4 | 2 | 6 | 4 | 1 | 3.4 |
Here's how it looks if I just pull the raw data. I want to recreate the column "Average Sale per Week", but only count from the week where the first sale occurred. So For row 1, that would mean to calculate (2+0+3)/3 = 1.6666667
For the second row, it would take all figures into account, even if the item didn't had sales for all weeks, it was still "active" in our assortment. (5+6+7+0+0)/5 = 3.6
The third row, should account for all weeks, since it had sales on each separate week.
In my application the weeks are labeled "YearWeek" and Item Numbers are labeled "itemno". The units, or data itself, is labeled "Fsgantal".
I hope this helps, sorry I can't post any thought out formulas yet, I haven't gotten that far.
