>>Also recently Hugo found an interesting case when IN subquery behaved differently than EXISTS because of NULLs involved.
>
>Actually I did not find anything, I just failed miserably to read the manual, as the behavior is documented and correctly pointed to me by Sergey. I am still embarrased by the whole incident, thanks for the reminder :)

No, it was not that obvious - I think it was great you posted that thread.