> > This looks bad to me, especially because you have a join going on in the
> > update. In fact, the comment clearly shows a false assertion, that ther
> > is only one relation in UPDATE.
> >
> > Is the update rewrite code assuming that the resdomno of an updated
> > column must match the attribute number? And the join is messing this
> > up?
> >
> > --
> > Bruce Momjian | 830 Blythe Avenue
>
> Right! The rewrite code assumes that the resdomno of the
> updated columns match the attribute number in the target
> relation. I don't know if the join is messing it up - but
> looks like. Thanks for the help - I think I have to look for
> usage of p_last_resno to find all the places where this can
> happen.
OK, do you need me to supply a patch? I would be glad to review
anything you have.
>
> Little joke:
>
> At the place in analyze.c, where a TLE is created, there is a
> comment that not creating a proper target list with correct
> resdomno's might break rules :-)
>
> I love those comments. And especially all that have XXXXX
> somewhere.
Yes, I have a few favorites.
--
Bruce Momjian | 830 Blythe Avenue
maillist@candle.pha.pa.us | Drexel Hill, Pennsylvania 19026
+ If your life is a hard drive, | (610) 353-9879(w)
+ Christ can be your backup. | (610) 853-3000(h)