Am I headed in the right direction? (long) - Mailing list pgsql-general

From Alan Young
Subject Am I headed in the right direction? (long)
Date
Msg-id 01051620082900.03112@harleypig.idiglobal.com
Whole thread Raw
Responses Re: Am I headed in the right direction? (long)  (will trillich <will@serensoft.com>)
List pgsql-general
I have the following structure:

create table _base (
 abbreviation varchar(10),
 short_name varchar(30),
 full_name varchar(60)
);

create table tome ( id serial ) inherits ( _base );
create table book ( id serial ) inherits ( _base );
create table chapter ( id serial ) inherits ( _base );

create table tome_book (
 tome_id int4 references tome ( id ),
 book_id int4 references book ( id )
);

create table book_chapter (
 book_id int4 references book ( id ),
 chapter_id int4 references chapter ( id )
);

create table verse (
 id serial,
 chapter_id int4 references chapter ( id ),
 number varchar(10),
 body text
);

create unique index tome_abbr_idx on tome ( abbreviation );
create unique index book_abbr_idx on book ( abbreviation );
create unique index tome_book_idx on tome_book ( tome_id, book_id );
create unique index chapter_abbr_idx on chapter ( abbreviation );
create unique index book_chapter_idx on book_chapter ( book_id, chapter_id );

create view verses as
select t.abbreviation as tome_abbr, t.short_name as tome_short, t.full_name as tome_full,
 b.abbreviation as book_abbr, b.short_name as book_short, b.full_name as book_full,
 c.abbreviation as chap_abbr, c.short_name as chap_short, c.full_name as chap_full,
 v.number, v.body
from tome t, tome_book tb, book b, book_chapter bc, chapter c, verse v
where (tb.tome_id=t.id)
 and  (tb.book_id=b.id)
 and  (bc.book_id=b.id)
 and  (bc.chapter_id=c.id)
 and  (v.chapter_id=c.id);

create rule insert_verses as
 on insert to verses
 do instead (
  insert into tome ( abbreviation, short_name, full_name )
  values ( new.tome_abbr, new.tome_short, new.tome_full );

  insert into book ( abbreviation, short_name, full_name )
  values ( new.book_abbr, new.book_short, new.book_full );

  insert into tome_book ( tome_id, book_id )
  values ( currval('tome_id_seq'), currval('book_id_seq') );

  insert into chapter ( abbreviation, short_name, full_name )
  values ( new.chap_abbr, new.chap_short, new.chap_full );

  insert into book_chapter ( book_id, chapter_id )
  values ( currval('book_id_seq'), currval('chapter_id_seq') );

  insert into verse ( chapter_id, number, body )
  values ( currval('chapter_id_seq'), new.number, new.body );
);

I would like to be able to ignore duplicate key entries and continue with the
remainder of the insertion.

For example, if tome already exists then I'll get a 'can't insert duplicate
value' error.  I'd like to ignore it and just add the remaining book, chapter
and verse values as well as adding the tome_book and book_chapter entries.

I'm kind of stumped ... any ideas?

Thanks.

Alan

pgsql-general by date:

Previous
From: "jacky_shu"
Date:
Subject: NEED Postgresql ODBC driver.
Next
From: bscinc2000@yahoo.comNOSPAM (Al Frick)
Date:
Subject: Re: Which Front End for Postgresql