I have a semantic dataset that reads data from a couple of dataflows upstream. These dataflows have a field called "LAST_REFRESH_DATETIME" that contains the timestamp when the dataflow was last refreshed.
I wanted to check that the "LAST_REFRESH_DATETIME" value was today and if it's not, to force the semantic dataset refresh to fail.
I have tried a few different methods where when I tested it on Power BI Desktop, it would fail immediately but when refreshing in Power BI Service, it would either completed successfully or just hang.
First method: Add a new column that checks if the "LAST_REFRESH_DATETIME" is >= TODAY(), if not, error. This method fails in Power BI Desktop with row level errors but ran successfully in Power BI Service 🤔
Second method: Validate the "LAST_REFRESH_DATETIME" date values and sum up the error flags. If the sum is greater than 0 then force the entire query to error out. This method fails to load the query in Power BI Desktop and hangs indefinitely in Power BI Service. I've tried waiting for over 20 mins and it just remain in In Progress status. I've then tried to go to the dataflow upstream and change the "LAST_REFRESH_DATETIME" value back into range and then lo and behold, the semantic dataset refresh that was hanging then completes within a min.
Has anyone encounter this issue, and if yes, have you found a way to force Power BI Service to fail a data refresh? From what I can tell based on the second method, it looks like Power BI Service is deferring the query evaluation when it encounter the error and seems to be just waiting for the upstream dataflow's issue to fix itself, though I'm not sure how often it's retrying the data refresh. This is not running in Fabric yet and I wouldn't want to move this to Fabric if this is going to be the behaviour!