Re: The folding of unquoted names to lower case in PostgreSQL

Поиск
Список
Период
Сортировка
От scott.marlowe
Тема Re: The folding of unquoted names to lower case in PostgreSQL
Дата
Msg-id Pine.LNX.4.33.0303181117540.5388-100000@css120.ihs.com
обсуждение исходный текст
Ответ на Re: The folding of unquoted names to lower case in PostgreSQL  (Tom Lane <tgl@sss.pgh.pa.us>)
Список pgsql-general
On Fri, 14 Mar 2003, Tom Lane wrote:

> "scott.marlowe" <scott.marlowe@ihs.com> writes:
> > Now that catalogs live in the
> > pg_catalog schema, I could just treat anything in the pg_catalog to be
> > fold to lower, while everything else would fold to upper.
>
> And you will determine whether something is in pg_catalog how, when you
> haven't yet done a catalog lookup for it?
>
> Keep in mind that the problem exists not only for system catalog names,
> but for column names in those catalogs, not to mention built-in
> functions.  So looking at whether the name starts with "pg_" really
> doesn't get you far as a way of deciding which way to fold.
>
> Possibly you could make something of "fold to upper case originally,
> but refold to lower case before looking in pg_catalog".  I have no idea
> how to implement that in a way that's not a horrid kluge though.  Also,
> it'd likely have unpleasant failure modes in some non-ASCII locales
> where upcasing and downcasing aren't quite inverses.  (Check the
> archives for past problems with keywords in Turkish...)

Agreed, it's a horrible kludge that way.  Is it possible to make it fold
everything to upper case, then do some kind of import/export of the system
catalog to match?  Or are parts of the catalog case sensitive, demanding
lower case?

For what we wanna do with it, a postgresql that folds everything to upper
is just fine.


В списке pgsql-general по дате отправления:

Предыдущее
От: Kurt Overberg
Дата:
Сообщение: Big insert/delete memory problems
Следующее
От: Matthew Phillips
Дата:
Сообщение: unexpected join results