Looks like it came back from the dead, and it's just as nasty as ever.  Maybe worse.

It is an obsolete syntax for doing a left join.

Chris, you need more conditions in your join to get all possible overlaps.

on
 ( t1.dtDATESTART BETWEEN t2.dtDATESTART and t2.dtDATEEND )  or
 ( t1.dtDATEEND BETWEEN t2.dtDATESTART and t2.dtDATEEND )  or
 ( t1.dtDATESTART < t2.dtDATESTART and t1.dtDATEEND...

The name of the article should be, “Get The DBA So Pissed Off At You That He Takes Away Your Access To Everything”, not “Let’s Play Block The DBA”, since...

It's better to just store all your data in a single column in a single row in a single table using the XML data type (2005 only).

It's the ultimate in...

Do you means the Windows Application Event log?  If so, that is what I would expect.

The Application Event log is not tied to SQL Server, so deleteing a job should...

Not to make too big a deal about it, I think a function for something like this has its place:

You may just need to call it once in a query...

I did quite a bit of testing of the performance difference of inline code vs. UDFs and posted it on this thread:

Demo Performance Penalty of User Defined Functions

http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=78601

One comment...

Geek fight!

It reminds me of this quote from Henry Kissinger:

"University politics are vicious precisely because the stakes are so small."

No one has posted on that thread for a couple of days, so maybe the fire is burnt out.

On another note: Is posting on this forum the best way to report...

I have done extensive testing against tables with millions of rows, and it is faster.

It is simple, it's much faster than converting to varchar, and it doesn't have any dependencies on the setting of DATEFORMAT or LANGUAGE.

The dateadd/datediff method show below is the best way to do it.

insert into table1 ( change_date )
select
 dateadd(dd,datediff(dd,0,change_date),0)
from
 table2

This will convert the date shipped to the first day of the month at time 00:00:00.000

dateadd(mm,datediff(mm,0,s.date_shipped),0)

As for the run time, we really don't have enough information about your tables, indexes, constraints, amount...

It is a bad design to store a date in a non-datetime column.

If you are going to change it to anything, change it to a data type of datetime.