Make it a binary index, and you will never regret it.

Index on deleted() tag deleted binary

An even better solution is to add a logical field called resuable, and set it to true instead of deleting the record. Then whenever you need to add a record, you update a reusable record instead. If no reusable record is available then you add a new record. This is how "real" SQL servers work, AFAIK.

>Hi All,
>
>What were the recommendations about creating an index on DELETED() records? Given a file of, say, 1 million records of which 10% might be deleted is it a good or bad idea to index on DELETED()?
>
>TIA