Andrew - Supernews wrote:
> On 2006-12-06, Timasmith <timasmith@hotmail.com> wrote:
> > Andrew - Supernews wrote:
> >> On 2006-12-06, Timasmith <timasmith@hotmail.com> wrote:
> >> > While pg_catalog.pg_index has the create index script I otherwise cant
> >> > find the index columns in the information_schema.
> >>
> >> That's because there are no index columns in the information_schema.
> >> Indexes simply do not exist in SQL (they are merely an implementation
> >> detail) and therefore are not included in information_schema (which is
> >> defined by the SQL spec).
> >
> > So because it is not in the SQL spec you think that is a good reason
> > not to have indexes?
> >
> > Or you think that we do not need that information in order to compare
> > to databases and create a DDL script to sync the two?
>
> You seem to be mistaking the information_schema for something which
> provides complete metadata - it does not, and due to the restrictions of
> the SQL spec which defines it, it never can be complete.
>
> --
> Andrew, Supernews
> http://www.supernews.com - individual and corporate NNTP services
never is a long time. Never ever? Never ever can any database
administrator every know which columns are indexed by the database?
We can never ever issue any advice as to which columns are best to
query for?
We can only ever for the forseable future pull up the string create
index definition and interpret it manually or with a flawed regular
expression?
Nah, I dont buy it.
I didnt buy Postgresql either of course.
But I still love it, I will do a parsing script for now.
