> I think that the required condition for all these ordering
Careless wording. the 'required' condition above means 'necessary
(but not sufficient)' condition so I think the lack of the
precondition (=multiple rows) results in prevention of the
postcondition = ordering problems.
> problems is generating multiple rows for single input (for a
> value of any column of the same table).
>
> If a prefixing set of values correspond to a unique index appears
> only once in a result, the result must can be fully-ordered by
> the extended pathkeys. This is what this patch stands on. It
> never be broken while the precondition is satisfied... I think.
>
> On the other hand, the precondition should be satisfied when all
> extended columns are simple Vars of the target table. I believe
> Vars cannot produce multiple result. More close checking for
> every possibility could be done but it should be a overdone.
regards,
--
Kyotaro Horiguchi
NTT Open Source Software Center