Thanks Gail,
Every time when google doesn't works for me, you have an answer 🙂
I got you point and tested it by examining the query plans.
Just a last clarification,
as u stated that for a unique non clustered index drop_existing makes sure that no duplicate value gets inserted during the time interval of Drop and create.
But drop_existing for a nonunique Non Clustered index should not have any effect.
as we are not bothered about duplicates in this case.