Currently, it's very easy to accidentally drop data sensitivity classifications when loading a table through CTAS, rename, and drop. If you load a table in this manner and you have data sensitivity classifications, you need to copy them over from the old table to the new before dropping the old table. This is a rather tedious manual effort. This should be made much easier, or the issue should be resolved in some other manner where the classifications are not dropped at all. Otherwise, we won't be able to trust an audit of that information. For an explanatory demo, see https://github.com/mlongoria/AzureSQLDW/blob/master/Azure%20SQL%20DW%20Data%20Classifications.ipynb