>Actually, unless I misunderstood your original question, both syntaxes should generate the same results
>
>But on the optimization, you're right, mine only does an index scan, whereas the one Naomi gave ( LIKE 'S[_]SH%' ) does an index seek and is better. I have a test table with 27 million names, mine took 1.5 seconds, Naomi's took .4 seconds.

That is a very interesting benchmark.

As for the fact that it should have returned the same result, I can't tell. Maybe there is setting locally that affects that. But, based on Naomi response, it seemed to be normal.