Jeff, I agree with you. Although I mentioned Mr Celko's name, in doing so I didn't intend to imply that I was taking his strict "ANSI good, proprietary bad" line. My view is that one should use ANSI SQL unless a proprietary syntax adds something in terms of performance or functionality. In terms of UPDATE...FROM, I think the opposite is true - that it's so dangerous that, in my opinion, it should only be used as a last resort. When I say "last resort", I mean if it can be shown that its performance is significantly better than the alternatives. In such circumstances, it should be thoroughly commented so that anyone reading the code knows that there is no possibility of cardinality errors, and no poor unsuspecting and inexperienced developers are tempted to use it elsewhere without due consideration. As always, and I should have said this in my previous post, test thoroughly before deciding which construction to use.