Re: Double sequence increase on single insert with RULE on

From: Sarunas Krisciukaitis <sarunask(at)lonus-tech(dot)com>
To: Michael Fuhr <mike(at)fuhr(dot)org>
Cc: Tomas Zerolo <tomas(at)tuxteam(dot)de>, pgsql-bugs(at)postgresql(dot)org
Subject: Re: Double sequence increase on single insert with RULE on
Date: 2005-11-17 08:11:29
Message-ID: 437C3B31.30001@lonus-tech.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Ok :) Then I found the solution in this partical case:
CREATE RULE test1_on_insert AS ON INSERT TO test1 DO INSERT INTO
test_log1 (qid) VALUES ( (SELECT lastval()) );
With this rule all inserts are working as expected :)

Thank you for you advise :)

Sarunas

Michael Fuhr wrote:

>On Wed, Nov 16, 2005 at 10:31:10AM +0200, Sarunas Krisciukaitis wrote:
>
>
>>I understand that RULES are like macros.
>>Strangest thing here is that INSERT to test1 will touch only one
>>sequence: test1_id_seq.
>>And it increments test1_id_seq twice during insert with RULE.
>>
>>
>
>Yes, that's a well-known effect of rewriting a query that includes
>a call to nextval(). NEW.id in the rule doesn't refer to the value
>that's inserted, but rather to the expression that's evaluated to
>get that value. Since you didn't provide a value for id it gets
>the default: nextval('test1_id_seq'). That expression is used in
>both inserts, so the sequence gets incremented twice. See the
>archives for numerous past discussions of this behavior.
>
>
>

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message Matt Carter 2005-11-17 08:27:00 Huge query stalls at PARSE/BIND stage (2)
Previous Message Acure 2005-11-17 00:31:50 BUG #2048: initdb hang up