I have a transaction table, and have to calculate distances to ID-s in time intervall e.g. hours into a new column like the unnamed one in this sample.
The problem is, that the first fact data can be 0, so first have to fill up the value by the last before maximum value of fact. If not avaible value have to find the last near equal value into fact.
So first have to fill up the fact by coulmn value whit a logic, and after that have to calculate the diference between fact values (grouped by ID and hour.)
Pls if u can help me
| ID | Time | Fact | |
| 7997 | 2019-01-03 10:13:00.000 | 159269,00 | 0,00 |
| 7997 | 2019-01-03 10:13:00.000 | 159269,00 | 0,00 |
| 7997 | 2019-01-03 10:13:00.000 | 159269,00 | 0,00 |
| 7997 | 2019-01-03 10:13:00.000 | 0,00 | 0,00 |
| 7997 | 2019-01-03 10:17:00.000 | 159269,00 | 0,00 |
| 7997 | 2019-01-03 10:57:00.000 | 159285,18 | 16,18 |
| 7997 | 2019-01-03 10:57:00.000 | 159285,18 | 0,00 |
| 7997 | 2019-01-03 11:06:00.000 | 159286,43 | 1,25 |
| 8001 | 2019-01-23 12:01:00.000 | 159986,68 | 0,00 |
| 8001 | 2019-01-23 12:01:00.000 | 159986,68 | 0,00 |
| 8001 | 2019-01-23 12:54:00.000 | 160008,04 | 21,36 |
| 8001 | 2019-01-23 13:05:00.000 | 160008,04 | 0,00 |
| 1214 | 2019-01-23 13:05:00.000 | 0,00 | 0,00 |
| 1214 | 2019-01-23 13:42:00.000 | 160013,29 | 0,00 |
| 1214 | 2019-01-23 13:44:00.000 | 160013,29 | 0,00 |
| 1214 | 2019-01-24 09:37:00.000 | 160024,82 | 11,53 |
| 3333 | 2019-01-24 09:39:00.000 | 160024,95 | 0,00 |
| 3333 | 2019-01-24 09:39:00.000 | 0,00 | 0,00 |
| 3333 | 2019-01-24 11:36:00.000 | 160036,16 | 11,21 |
| 3333 | 2019-01-24 15:24:00.000 | 160081,58 | 45,42 |
(distinctcount ID more than 10000, the whole data more than 1m rows)
