We have been using Power Platform dataflows to bring data from Snowflake into Dataverse tables. Recently, we had an issue where the dataflow thought the query returned zero rows. Since our dataflow was set up to delete any rows that did not exist in the query, all of our data was deleted for this table and the relationships with several other tables were broken. The next time the dataflow ran, it repopulated the table because the query returned data.
Since this has only happened the one time, I do not believe this was an issue on the Snowflake side. Instead, I suspect that the dataflow couldn't connect to Snowflake for some reason, and instead of handling that as an error, it handled it as returning zero rows.
Can anyone confirm that if a connection cannot be made within a dataflow, it throws an error instead of proceeding as though it returned zero rows?
We'd love to be able to use this functionality as it was intended, but right now, it's just too risky and would cause too much disruption if this issue happened again.