Re: text fields and performance for ETL

Поиск
Список
Период
Сортировка
От Laurenz Albe
Тема Re: text fields and performance for ETL
Дата
Msg-id 12598e4d44a8bf530934e793fde9ab2743108046.camel@cybertec.at
обсуждение исходный текст
Ответ на Re: text fields and performance for ETL  (Bruce Momjian <bruce@momjian.us>)
Ответы RE: text fields and performance for ETL
Список pgsql-docs
On Fri, 2021-11-05 at 11:27 -0400, Bruce Momjian wrote:
> On Fri, Nov  5, 2021 at 07:32:12AM -0700, David G. Johnston wrote:
> > On Friday, November 5, 2021, Bruce Momjian <bruce@momjian.us> wrote:
> > 
> > 
> >     >
> >     > Perhaps, right before the tip you quoted, something like that:
> >     >
> >     >   If your use case requires a length limit on character data, or
> >     compliance
> >     >   with the SQL standard is important, use "character varying".
> >     >   Otherwise, you are usually better off with "text".
> > 
> >     I can support that if others think it is valuable.
> > 
> > 
> > 
> > The motivating complaint is that we should be encouraging people to use varchar
> > (4000) instead of text so external tools can optimize.  If we are not going to
> > do that I really don’t see the pointing in changing away from out current
> > position of “only use text”.  True length limit requirements for data are rare,
> > and better done in constraints along with all other the other constraint that
> > may exist for the data.  I believe comments with respect to the SQL standard
> > are already present and adequate.
> 
> Agreed.

+1, so let's leave it as it is.

Yours,
Laurenz Albe




В списке pgsql-docs по дате отправления:

Предыдущее
От: Bruce Momjian
Дата:
Сообщение: Re: text fields and performance for ETL
Следующее
От: PG Doc comments form
Дата:
Сообщение: Add link to unicode collation docs