@Hemant
If the order by clause provided yields a unique result, then we would get deterministic output.
For ex:
If we create a rowID for this dataset, with CustomerID used in OrderBy clause, then depending upon the runtime, we may get non-deterministic output for CustomerID 100, because any record may get picked up for Row Number 1.
| CustomerID | SalesID | OrderDateTime |
| 100 | 900 | 2023-07-08 10:00:00 |
| 100 | 901 | 2023-07-09 10:00:00 |
| 101 | 902 | 2023-07-09 10:00:00 |
In cases like this, it is recommended to use a column to define the ordering in a deterministic way. The usage of OrderDateTime column in the OrderBy clause will help us to achieve a deterministic output.
The same is also documented in the official documentation - https://docs.databricks.com/sql/language-manual/functions/row_number.html#:~:text=If%20the%20order%2....
