On Fri, May 17, 2019 at 10:21 AM Kyotaro HORIGUCHI
<horiguchi.kyotaro@lab.ntt.co.jp> wrote:
>
> We now have several syntax elements seemingly the same but behave
> different way.
>
> At Thu, 16 May 2019 15:29:36 -0400, Robert Haas <robertmhaas@gmail.com> wrote in
<CA+TgmobK1ngid9Pxs7g8RFQDH+O1X4yyL+vMQtaV7i6m-Xn0rw@mail.gmail.com>
> > On Thu, May 16, 2019 at 2:56 PM Fujii Masao <masao.fujii@gmail.com> wrote:
> > > Yes. Thanks for the comment!
> > > Attached is the updated version of the patch.
> > > It adds such common rule.
> >
> > I'm not sure how much value it really has to define
> > opt_boolean_or_string_or_numeric. It saves 1 line of code in each of
> > 3 places, but costs 6 lines of code to have it.
>
> ANALYZE (options) desn't accept 1/0 but only accepts true/false
> or on/off. Why are we going to make VACUUM differently?
>
> And the documentation for ANALYZE doesn't mention the change.
Commit 41b54ba78e seems to affect also ANALYZE syntax.
If it's intentional, IMO we should apply the attached patch.
Thought?
Regards,
--
Fujii Masao