Sorry for the late reply but it's been a very busy week for me. Finally got time to look at these posts and it's always great to see people thinking...

Have a look at this article which I wrote about this question and a way of getting the results using binary flags that performs well with many rows.

http://www.sqlservercentral.com/articles/T-SQL/107611

Submitted and available for your perusal here.

http://www.sqlservercentral.com/articles/T-SQL/107611/

It's not available to buy yet but if you go to the website and click the contact us link then ask to be an alpha tester you might get to...

That would indeed be much faster and simpler if it wasn't for the possibility of duplicate checksums.

This link http://ask.sqlservercentral.com/questions/7253/limitations-w-using-checksum-for-comparing-data-wh.html mentions this very problem and the only way around...

Oh dear. Would have gone for the correct answer if I hadn't paused and asked myself what this question really wanted. My interpretation was that you can't update statistics on...