The only glitch is that the response IDs are not conserved (the new ones are follows the old ones).

I don't understand this part. What do you mean with the new ones follow the old ones? For me this sounds about right, that the new responses that you collect after the reactivation start after the ones that you re-imported. Or did I understand you wrong?

Before the deactivation, there was 36 responses whose IDs range fro m1 to 36. After re-import the IDs of the imported responses range from 37 to 72 (there is no new responses for now).

Just to be complete: I did not reimported the responses from the CSV file; I used the builtin functionality to "re-use" the archived response table.

Yes,

This is the issue with get from old table … i prefer to use VV for this issue.

OK. Not a great problem but I will keep it in mind for the next time.

Before the deactivation, there was 36 responses whose IDs range fro m1 to 36. After re-import the IDs of the imported responses range from 37 to 72 (there is no new responses for now).

OK, now I understand. I don't think it is a big deal, but yes, it would be nicer if they would come back with their original IDs, if you use the ID for example to identify them.

I think it can be a great feature request for DB backup ?
No ?