Re: [SQL] Rules and sequences

2010-05-27 Thread A. Kretschmer
In response to Tom Lane :
 Ben Morrow b...@morrow.me.uk writes:
  I am trying to implement a fairly standard 'audit table' setup, but
  using rules instead of triggers (since it should be more efficient).
 
 Rules are sufficiently tricky that I would never, ever rely on them for
 auditing.  Use a simple AFTER trigger instead.

There are a ready solution:
http://andreas.scherbaum.la/blog/archives/100-Log-Table-Changes-in-PostgreSQL-with-tablelog.html


Regards, Andreas
-- 
Andreas Kretschmer
Kontakt:  Heynitz: 035242/47150,   D1: 0160/7141639 (mehr: - Header)
GnuPG: 0x31720C99, 1006 CCB4 A326 1D42 6431  2EB0 389D 1DC2 3172 0C99

-- 
Sent via pgsql-sql mailing list (pgsql-sql@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-sql


Re: [SQL] Rules and sequences

2010-05-27 Thread Ben Morrow
Quoth t...@sss.pgh.pa.us (Tom Lane):
 Ben Morrow b...@morrow.me.uk writes:
  I am trying to implement a fairly standard 'audit table' setup, but
  using rules instead of triggers (since it should be more efficient).
 
 Rules are sufficiently tricky that I would never, ever rely on them for
 auditing.  Use a simple AFTER trigger instead.

OK, thanks. 

Ben


-- 
Sent via pgsql-sql mailing list (pgsql-sql@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-sql


[SQL] Rules and sequences

2010-05-26 Thread Ben Morrow
I am trying to implement a fairly standard 'audit table' setup, but
using rules instead of triggers (since it should be more efficient).
However, I'm running into problems when one of the audited tables has a
'serial' column that is allowed to default:

create table foo (id serial, bar text);
create table audit (ix bigserial, rec text);
create rule audit_insert as on insert to foo do also 
insert into audit (rec) values ((new.*)::text);
insert into foo (bar) values ('baz');
select * from foo;
 id | bar 
+-
  1 | baz
(1 row)
select * from audit;
 ix |   rec   
+-
  1 | (2,baz)
(1 row)

I can see why this is happening (the rule is essentially a macro, so the
NEW expression gets expanded twice, including the nextval call, so the
sequence is incremented twice), but is there any way to prevent it? Some
way of 'materialising' the NEW row so it is just plain values rather
than a list of expressions?

Ben


-- 
Sent via pgsql-sql mailing list (pgsql-sql@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-sql


Re: [SQL] Rules and sequences

2010-05-26 Thread Tom Lane
Ben Morrow b...@morrow.me.uk writes:
 I am trying to implement a fairly standard 'audit table' setup, but
 using rules instead of triggers (since it should be more efficient).

Rules are sufficiently tricky that I would never, ever rely on them for
auditing.  Use a simple AFTER trigger instead.

regards, tom lane

-- 
Sent via pgsql-sql mailing list (pgsql-sql@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-sql