I completely agree. The data source we're using for some of our dataflows does not have a "record created on" field, so we're using a fiscal period start date field for incremental refresh instead. The problem is that the data contains forecast information for future periods, so incremental refresh doesn't pick this up. The workaround we're using is to have one query for current and past periods, another query for future periods, and a third query to append the first two. We setup incremental refresh only on the current/past periods query. The overall refresh time is a little faster, but not by much, because the append query takes additional time to run. Another downside is that we have more than one query set to "enable load", and the users can get confused when selecting the dataflow to pull into their datasets.
Incremental Refresh allows you to use a DirectQuery partition for future data. It would be helpful to be able to configure that for use with import mode, too. I have a large fact table that inc...
