Toby Harman (7/5/2011)
Good article covering a common situation. Indexes are created because we assume they will be the access path (as pointed out in the DateTime discussion) but because they either aren't covering the entire query result or the result set is too large, we end up with indexes that are never used.As to the "It will get slow if we drop these indexes" discussion, my only comment to that is "We can add them back". It's not like they purging data which cannot be recovered.
The moment they are dropped, users will believe the system is running slower. Furthermore any application problems will be initially blamed on the missing indexes.