Andreas Joseph Krogh wrote:
When in doubt, consult the standard ... Oracle's treatment of NULL is
known to violate the standard, IIRC. Your measure of correctness seems
to be "appears to me more logical", but ours is "complies with the
standard".
I know PG violates the standard in other places and core's favourite argument
for doing so is "the standard is braindead here, so we do it our way".
In very few places. If you think that an argument like that will make us
break well established standards-compliant behaviour, you are surely
sadly mistaken.
cheers
andrew
---------------------------(end of broadcast)---------------------------
TIP 9: In versions below 8.0, the planner will ignore your desire to
choose an index scan if your joining column's datatypes do not
match
