All screenshots were taken from an RDC on the server.

Machine is 64bit, and this package has "Run64BitRuntime" set to False (because I've never connected to a DBF successfully in 64bit).

I...

Stick with me Kraig...

I'm attaching a screeny of the resource monitor's take on memory which is why I originally said it's not taxed.

Also, I saw on the "output" tab, this...

fast-load... never saw it have such a dramatic effect.

SORT in SSIS... yeah, learned that pretty quickly. I've ORDERED BY at the source to overcome this.

Lookup/FullCache... didn't know sorting was...

Even though the slow down appeared to be at the merge join in those PNG's, I changed the OLE DB destinations to "Rowset with Fast Load", and got a 100:1...

response marked as answer. Much obliged fellas.

Yeah, agreed. Didn't know if I was missing something obvious. I'm going to create an actual table to this end and simplify my life. Moden would be...

I'm *trying* to use it as a table in that statement. It's purpose it to simply limit the rows affected in the UPDATE. It's getting populated inside a...

Nah, the 400MM will digest nicely if it's done set-based. I won't loop through it. The recordset in question has only a single column of integers; [id]'s of...

400MM rows

Thanks for the nods fellas. I'm headed down that path right now.

Correct me if I'm wrong, but if the function that JLK proposed were to be inserted into an otherwise set-based UPDATE or SELECT statement, would it not then be executed...

Thanks Phil. That's what I needed.

Thanks for setting me straight. Not sure why I was convinced otherwise.

It just seems like a workaround to me. I'd rather trim the fat and get 4 part naming to work as it should.

Is there a reason why referencing another...

BTW, all SQL Servers in the domain are running in Mixed Mode authentication, and I'm logged in as a domain administrator. Thanks!