I thought it was more correctly we were considering not using the the
system locale automatically, but that if someone wished to use
--locale=en_US we'd let that work, right?
I would assume that if someone actually went to the bother of setting a
locale, then it should be the deciding factor in how we handle dates, et.
al.
On Tue, 24 Jun 2003, Bruce Momjian wrote:
>
> We are actually considering not honoring locale for initdb encodings, so
> it might make no sense to do this --- that another reason for the
> question mark, but until we decide, it is an open issue.
>
> ---------------------------------------------------------------------------
>
> Lincoln Yeoh wrote:
> > At 03:24 PM 6/23/2003 -0400, Bruce Momjian wrote:
> >
> > >Added to TODO, with question mark:
> > >
> > > * Have initdb set DateStyle based on locale?
> >
> > Given various issues with locale (indexes, ordering etc) I'd think that
> > having a DB follow the O/S locale should be special case and require
> > explicit configuration.
> >
> > More so if certain locales are significantly slower than others which
> > seemed to be the case at least in recent memory.
> >
> > What if a European DB backed website is hosted on a US server with English,
> > French and German data?
> >
> > If apps/programs are talking to DBs more than people are then it may make
> > more sense to store things in an application friendly format e.g. (date =
> > YYYY-MM-DD, or seconds since epoch) format and having the app convert it
> > based on the user's preferences. After all even in English, apps may choose
> > to display Tuesday as T, Tue, Tuesday, or whatever the Boss wants.
> >
> > Unless postgresql has special features allowing switching from one locale
> > to another on the fly (including indexes, ordering etc) within a DB
> > session, I'd rather stick to say the C locale, or whatever it is that's
> > fastest.
> >
> > Another point of consideration: if someone accidentally loads
> > multibyte/other locale data into a C locale DB (or whatever is chosen as
> > default DB locale), would dumping the loaded data and reloading it into a
> > multibyte locale result in information/precision loss?
> >
> > Link.
> >
> > ---------------------------(end of broadcast)---------------------------
> > TIP 8: explain analyze is your friend
> >
>
>
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match
