Re: newbie needs help with postgres internal tables

Le Friday 23 November 2001 14:14, Tom Lane a écrit :
> >  It's necessary cause I want to display in a treeview ( GtkCtree ) all
> > the the tables in a particular database, and for each table, all the
> > fields in this table ).
>
> You can only see tables of the database you are connected to.  To do the
> above, you'd need to open a connection to each database in turn
> (assuming that the installation is configured to let you get into all of
> them...)
 Yes, and that's what I already do, but I always see the tables of this
database AND postgresql ( pg_* ) tables.
>
> pg_database is visible from all databases, and so are pg_shadow and
> pg_group, but AFAIR everything else is database-local.

  See here the result of select in psql while connected to gpf database:

gpf=> select * from pg_tables;
   tablename    | tableowner | hasindexes | hasrules | hastriggers
----------------+------------+------------+----------+-------------
 pg_type        | postgres   | t          | f        | f
 pg_attribute   | postgres   | t          | f        | f
 pg_proc        | postgres   | t          | f        | f
 pg_class       | postgres   | t          | f        | f
 pg_group       | postgres   | t          | f        | f
 pg_database    | postgres   | f          | f        | f
 pg_variable    | postgres   | f          | f        | f
 pg_log         | postgres   | f          | f        | f
 pg_xactlock    | postgres   | f          | f        | f
 pg_attrdef     | postgres   | t          | f        | f
 pg_relcheck    | postgres   | t          | f        | f
 pg_trigger     | postgres   | t          | f        | f
 pg_inherits    | postgres   | t          | f        | f
 pg_index       | postgres   | t          | f        | f
 pg_statistic   | postgres   | t          | f        | f
 pg_operator    | postgres   | t          | f        | f
 pg_opclass     | postgres   | t          | f        | f
 pg_am          | postgres   | t          | f        | f
 pg_amop        | postgres   | t          | f        | f
 pg_amproc      | postgres   | f          | f        | f
 pg_language    | postgres   | t          | f        | f
 pg_aggregate   | postgres   | t          | f        | f
 pg_ipl         | postgres   | f          | f        | f
 pg_inheritproc | postgres   | f          | f        | f
 pg_rewrite     | postgres   | t          | f        | f
 pg_listener    | postgres   | t          | f        | f
 pg_description | postgres   | t          | f        | f
 pg_shadow      | postgres   | f          | f        | t
 gpf_db         | gpf        | t          | f        | f
(29 rows)


 For information I finally found two solutions :

1- Maintain myself a table containing the tables of each database.

2- Automatically naming all the tables of a database using the same rules
  ( ie : gpf database, all the tables are named gpf_*), so that a
  "select * from pg_tables where tablename like 'gpf%' " will return what I
   need.
    I just have to keep the naming convetion of each database, which is not a
    problem since I already have one internal table containing a row row for
    each database created and ruled by GPF ( my future application ).

 I chose the second solution ( probably faster and easier ).

 Thanks for your answers.

 PS: Sorry if my english is not always the best, I'm French ;-)

--
cordialement,  William WAISSE fpr = 9CCD 7DA2 7050 8805 F471  03D1 DF76 B78C
690B 4E07
  --
Computers are like air conditionners. They work better when you close windows.
  --
Visitez donc mon site perso (version 0.4 du 29/05/2001) :
http://www.neofutur.net
  --
Vous habitez dans l'Essonne et vous aimez les Pingouins ???
visitez donc la GAULE(=LUG91) : http://www.gaule.org
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCS/E/MU d- s: a- C++++ UL+++ P+ L++++ E--- W++ N o-- K- w---
O- M- V- PS+ PE-- Y++ PGP++ t+ 5 X++ R++ tv- b+++ DI- D G e++ h* r++ y+++
------END GEEK CODE BLOCK------