Hi Community,
Accumulating Snapshots are a model-technique based on Kimball (Accumulating Snapshot Fact Tables | Kimball Group)
Essentially you have a row that is visited on each update. Normal transactions facts are inserts only, accumulating snapshots receive updates, usually on a date column.
I want to set up an incremental refresh and have followed the great article on Tabular Editors (What is a Refresh Policy? | Tabular Editor Documentation).
Imagine you have a sales-table with these columns:
orderId
dateKey@Order
dateKey@Picking
dateKey@Packing
dateKey@Outgoing
SalesAmount
I have an order dimension and a date dimension. I want to calculate several measures (role-playing e.g. SalesAmount Order Quantity or SalesAmount Pickunt Quantity etc.)
Now I want to use Incremental Refresh. But I cannot use just one column for RangeStart and RangeEnd. I want to partition over dateKey@Outgoing and also if dateKey@Outgoing is empty.
Did anyone solve incremental refresh with this modelling technique? I'd be very curious!
Thank you and best regards,
Konstantin
