SQLkiwi (8/4/2011)
Tom.Thomson (8/4/2011)
...if it delivers a non-zero result the next thing to happen will probably be the conversion and that was probably already done once...Indeed. Worse, people might write the test and the convert in the same statement in such a way that the convert could be reordered by the optimizer to happen before the test...
Yes indeed. I've seen similar things. I wouldn't have seen them if people didn't tend to think that things are evaluated in the "obvious" order, but this would be another chance for people to make that assumption and get burnt by it.
Tom