Ok, anyone up to a challenge? A co-worker and I are debating the following: I have a table that I want to ETL. Valid data goes to a "F00" table and invalid data to an "E00" table. The debate is which is faster? Using a stored proc that references a view to populate the "F00" table on no errors and the same view to populate the "E00" table on the error condition. Or...using one query in a DataTransform package with the data flow going to the "F00" and "E00" tables respestively on the error condition. The big question is this: We know that the view is set based. But what about the DataFlow Task? Is it set based or row based? In other words does a DataFlow Task implement a cursor beneath the sheets?