Rob, you're right ... that's what I get for doing a drive-by shooting ...*arg*
---------------
Hmmmm, and rolling this into a where/SARG clause it looks like the query engine does halt processing of rows once the time limit has been hit.
====================
-- let run to completion
select convert(varchar,dbo.f_getdate(so.id),109)
from sysobjects so, syscolumns
go
200K rows, cpu = 7500ms, logical IOs = 1,830
table/index accesses = ~400K each
-- let run for 2 seconds
select convert(varchar,dbo.f_getdate(so.id),109)
from sysobjects so, syscolumns
where dbo.f_getdate(so.id) <= dateadd(ss,2,getdate())
go
55K rows, cpu = 2000ms, logical IOs = 200
table/index accesses = ~110K each
====================
