Does this only optimize things like `select * from table where false` type stuff or can it optimize access paths as well? such as whether an index can be used? and what ranges? or should a sequential scan be used