A warning message in Power BI Desktop says "It is not recommended to use incremental refresh with non-foldable queries."
However, incremental refresh for a non-folding data source actually does seem to work. I'm wondering what the actual limitations are then for non-folding data sources?For example, if I have a data source that accepts start/end dates in the query string, and I simply directly substitute in those start/end parameters into the string:
data = MyDataSource("[...], dates=range(" & RangeStart & "," & RangeEnd & ")")
I can get it to only query for the past month upon subsequent refreshes after querying for six months of historical data the first time.
Even if I add some kind of simple transformation afterwards such as SelectRows, it still doesn't seem to break incremental refresh.
Cases like this seem to suit our needs just fine.
If it’s a pure import mode policy without real-time data, the query mashup engine might compensate and apply the filter locally, which requires retrieving all rows for the table from the data source.
But would this pitfall only apply if I try to apply SelectRows(..., each [start_data] >= RangeStart and [end_date] < RangeEnd) to do the date range filtering instead of the direct string substitution above? Or are there other things that can go wrong with non-folding data sources with incremental refresh?
