> . but I'd say the same thing about Postgres recommending against use of NOT IN, which is a common construct and more readable than an anti-join. The recommen

> . but I'd say the same thing about Postgres recommending against use of NOT IN, which is a common construct and more readable than an anti-join. The recommendation against that is due to the semantics defined in the standard. If there is any NULL in the list matched against, the entire a NOT IN (...) expression also has to evaluate to NULL. Which is extremely confusing to most people (and has some consequences in how efficient things can be executed).