davidr (9/4/2008)
side note: we shouldn't really be doing "select *" to check for existence. If that's a LOONNNNGGGGG row there could be a lot of overhead to no good purpose. It may be more self-documenting to code "select 'exists' from ....."
Hi David,
I a stand-alone SELECT statement, I would agree. But in the context of a [NOT] EXISTS subquey, it really doesn't matter a bit what you put behind the SELECT - whatever you write there, SQL Server will interpret it as checking for existance of a row, and perform the check in the most efficient way.
Personally, I prefer SELECT * as it indicates checking for a row, not checking for a specific value or so. But that's just personal preference. From a performance viewpoint, there is absolutely no difference.