That implies that if there is any logic change for complete tables since it is recomputed on every update it is safe to not refresh the pipeline, where as with incremental tables if there is any change to recompute the changes we should refresh the pipeline, right?