Обсуждение: [HACKERS] Clarification of NULL values

from X3H2-97-030 (I think)


            c) Let P be any row of TS and let Q be any other row of TS,
              and let PVi and QVi be the values of Ki in these rows,
              respectively. The relative position of rows P and Q in the
              result is determined by comparing PVi and QVi according to
              the rules of Subclause 8.2, "<comparison predicate>", where
              the <comp op> is the applicable <comp op> for Ki, with the
              following special treatment of null values. Whether a sort
              key value that is null is considered greater or less than
              a non-null value is implementation-defined, but all sort
              key values that are null shall either be considered greater
              than all non-null values or be considered less than all non-
              null values. PVi is said to precede QVi if the value of the
              <comparison predicate> "PVi <comp op> QVi" is true for the
              applicable <comp op>.


And if anyone can follow that.......

Adrian

------------------------------

adrian@hottub.org wrote:
> >From X3H2-97-030 (I think)
> ... with the
> following special treatment of null values. Whether a sort
> key value that is null is considered greater or less than
> a non-null value is implementation-defined, but all sort
> key values that are null shall either be considered greater
> than all non-null values or be considered less than all non-
> null values. PVi is said to precede QVi if the value of the
> <comparison predicate> "PVi <comp op> QVi" is true for the
> applicable <comp op>.

Well, it means that as long as the nulls aren't scattered in the middle
of the returned ordered values you can do anything you want :)

Postgres does it correctly according to this standard...

            - Tom

------------------------------