On Tue, 2002-04-30 at 03:35, Tom Lane wrote:
> Rod Taylor <rbt@zort.ca> writes:
> > Appears psql needs to know how to differentiate between it's own temp
> > tables and those of another connection.
>
> More generally, psql is as yet clueless about schemas.
>
> regression=# create schema foo;
> CREATE
> regression=# create schema bar;
> CREATE
> regression=# create table foo.tab1 (f1 int);
> CREATE
> regression=# create table bar.tab1 (f2 int);
> CREATE
> regression=# \d tab1
> Table "tab1"
> Column | Type | Modifiers
> --------+---------+-----------
> f1 | integer |
> f2 | integer |
>
> This is ... um ... wrong. I am not real sure what the right behavior
> is, however. Should \d accept patterns like schema.table (and how
> should its wildcard pattern matching fit with that?) If you don't
> specify a schema, should it only show tables visible in your search
> path?
Yes.
For me the intuitive answer would be
regression=# \d tab1 Table "foo.tab1" Column | Type | Modifiers--------+---------+----------- f1 |
integer|
Table "bar.tab1" Column | Type | Modifiers--------+---------+----------- f2 | integer |
i.e. default wildcarding of missing pieces
-------------
Hannu
