Incremental refresh - "Store rows" based on specific start date (instead of time period)?

[Anonymous]

Hi All,

Just wanted to confirm: when setting up an Incremental Refresh, it appears it's only possible to define the 'Store rows in the last' as a time period (e.g. 30 days, etc).

Is there any way to tell the incremental refresh 'Store rows' based on a specific start date instead, or change this value dynamically/automatically (e.g. days since date)?

E.g. if my data has a start date of 1 September 2021, I could set 'Store rows in the last' 30 days, but from tomorrow it will start purging data after the 1st.

I could increase this time period to a higher value - but has it's own problems, either:

- I have to remember to come back and change the value higher as it reaches time period, or

- results in unnecessary calls/refreshes to the data source for days worth of data that doesn't exist.

[lbendlin]

that's not how partitions work.  They are on Day, Month, Quarter and Year granularity level, and they only work with UTC datetime stamps and regular calendars (not fiscal calendars)

So when you say "Last 30 days" this will be translated by the partition engine and it will then decide which partitions to refresh, which to consolidate, and which to leave alone.

If you want more control you will want to read up on XMLA endpoints and MDX queries to manipulate individual partitions.

Advanced incremental refresh with the XMLA endpoint in Power BI - Power BI | Microsoft Docs