Vague question about deleting duplicate rows.
Sigiloso
Tried to clarify some details like whether the rows had timestamps or some kind of internal row id. The "answer" depends on the tools at hand. The best idea is to avoid the situation in the first place. While there are cursor approaches and CTE hacks that allow "fixes" I discussed set-based database theory as preferable from a general perspective rather than feeding her a stock "solution" to a complex problem that really warrants a deeper understanding of what created the problem in the first place and why the hacks aren't always available and come with their own risks. My bad for knowing more than the interviewer.