Forum Discussion
Power BI Incremental Refresh Behavior Question
- 1 year ago
Parameters for range start and range end applied to the last_updated_date in the tableLet me stop you right there. last_updated_date is not a suitable column for incremental refresh partition policies. This will result in data duplication across partitions. Instead use an immutable column like created_date.
Parameters for range start and range end applied to the last_updated_date in the table
Let me stop you right there. last_updated_date is not a suitable column for incremental refresh partition policies. This will result in data duplication across partitions. Instead use an immutable column like created_date.
LOL ok that is what I thought and that is what I tried to tell the BI team.
This leads me to my next question: if the parameters are applied to the created_date instead and the incremental refresh is still set to pull in 7 days and the detect data changes is unchecked, will power BI pull the data from the source by created_date?
What I can't figure out is how does Power BI know which data to replace? There's no key column defined so it can't be performing a merge.
- lbendlin1 year agoSuper User
There is no merge in any of this. You have partitions. Their queries are managed by the RangeStart and RangeEnd parameters which are controlled by the Power BI service based on your requirements*.
They are then appended.
* I shoud say "guidance. You may specify 7 days but the service will decide for itself how many partitions to actually refresh, which ones to consolidate etc.
https://learn.microsoft.com/en-us/power-bi/connect-data/incremental-refresh-xmla