Interesting stuff.

I have to ask though, why not just manually make this yourself with 2 tables and some triggers? All it's doing is tracking the history of the table into another table and adding additional syntax to go back into time to see the results of your data say, 6 months ago?

When I watched the video on this new feature, they didn't really touch on the subject of storage. This is basically copying the record through every change in the temporal table. For large tables, that's going to grow pretty large through each iteration of that record with a separate record and timestamp. The speaker kept saying things like, "Oh, nice! So, I don't have to worry about this table!" Kind of made my right eye twitch. 😎

I think some use cases for me is the slowly changing dimensions. Another would be for auditing the records through the cleaning process as I don't control the source data we are storing. So, being able to go back to the original raw record and track the iteration of that record during the transformation process could be useful too. I would have a copy of the original record in the flatfile that's compressed and archived after the importation, a copy of the cleaned record in the actual table and multiple copies of the original and cleaned record through the transformation process in the history table.