I think I am the same page as you now, sorry, I missed your *I* inference in your OP.

The intertubes did not turn much up for me, as I suspect...

The optimization issue is what keeps me away from table variables. I had heard about the recompile option but it did not sway my view on things given the expense...

One point for clarification:

The advisor is most useful when run well before your migration effort begins. It inspects your current system and provides information on...

My initial thoughts went towards explicit transactions and transaction isolation levels. What is the real-world scenario behind the question? And are you equally worried about locking on both tables, or...

srferson-957148 (6/25/2011)

create table #test

(Part...

@dji I sense a little frustration in your posts 🙂 It seems like you're "in it" and will be "in it" for a little while longer so I wanted to...

New temp tables will use the default collation of tempdb unless otherwise specified.

Here are a few options (there may be others) for you in terms of creating your temp tables...

I can vouch for GnuPG being solid. It's free unless you require the patented Idea algorithm in which case you'll need to buy a license for that.

Wow, it must be spring, err summer, cleaning time again. I think I have seen a thread like this 3 or 4 times this week. Another good suggestion from another...

An env var to point to a package config is a good start but it goes downhill from there. This is probably a .NET developer trying to treat the Script...

I think what you want is to use a CTE that makes use of the ROW_NUMBER() function PARTITION BY configurationname and ORDER BY enddate DESC, and then you will select...

I would like to help. Do you mind providing some DDL and DML to build the test environment consistent with your post?

Here is what I mean:

http://www.sqlservercentral.com/articles/Best+Practices/61537/%5B/url%5D

I ran the test script after adding the TABLOCK hint to the target # table on INSERT and the end result was similar: for heaps SELECT...INTO is still faster than...

The computed column IS a workaround. You are screwed by poor database design. The proper way out of the mess is to redesign the schema so columns that naturally need...

This is what I mean about adding a computed column and having it participate in an index:

ALTER TABLE OrderHeader ADD OrderHeader_ID2 AS (CAST(OrderHeader_ID AS VARCHAR(100))) ;

CREATE INDEX [ix_OrderHeader.OrderHeader_ID2] ON OrderHeader...