Incremental Refresh - Reducing "Load" on Datasource

[Johann_32]

Hello,

I wanted to understand the concept of incremental refresh and already did a test implementation and uploaded a dataset that has it configured, in our Power BI service.

The benefit of incremental refresh is, that,

- only data is loaded that is changing frequently

- only data is considered to be refreshed in a time interval that you are defining (RangeStart, RangeEnd).

But from a technical perspective, what does actually happen in the source system where the data is coming from?

Let's assume the data is coming from a SQL Database, is the stress/load now any higher or lower due to the incremental refresh?

I think the load on the database should decrease, as it only gets queried for few data, as where before it was always queried for the whole set of data.

Thank you.

Best Wishes

Johann

[lbendlin · Accepted solution]

What really happens is that Power BI Service stores the results of the query against your data source in separate partitions. Based on how you specify your parameters it will then only run subqueries for the partitions that need to be updated, and will not run any queries for older partitions.

This will reduce the load on your source system, but only if you have an index on the field that you use for RangeStart/RangeEnd.