pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Postgres DBA <postgres@nest.bistbn.com>
To: Marc Howard Zuckman <marc@fallon.classyad.com>
Cc: M Simms <grim@argh.demon.co.uk>
Cc: SQL Mailing List <pgsql-sql@postgreSQL.org>
Subject: Re: [SQL] abusing an aggregate funct
Date: Mon, 16 Nov 1998 10:31:45 +0200 (IST)
Message-ID: <Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com> (raw)
In-Reply-To: <Pine.LNX.4.02A.9811151101140.8582-100000@fallon.classyad.com>



On Sun, 15 Nov 1998, Marc Howard Zuckman wrote:

> On Sun, 15 Nov 1998, Postgres DBA wrote:
> 
> > 
> > 
> > On Sun, 15 Nov 1998, M Simms wrote:
> > 
> > > > Unfortunately, solution using sequences is not so good unless you don't 
> > > > suppose to use access to that table from some simultaneously running
> > > > sessions. The problem is that every sessions accessing this additional
> > > > sequence  with nextval()  will get its own pool of values for the
> > > > sequence, so early or later you'll get some gaps in records numbering
> > > > because of at least on of two reasons:
> > > 
> > > >From the manpage of create sequence
> > > 
> > > 
> > >        Low-level locking is used to enable multiple  simultaneous
> > >        calls to a generator.
> > > 
> > > Doesnt this mean it will work for this task?
> > > 
> > > 
> > 
> > Sure, you can access one sequence from a couple of sessions at once but
> > it the only thing is guaranteed is UNIQUENESS of values you get from
> > nestval() across one sequence. Although nobody can promise you that there
> > will no gaps... There are lots of reasons, I mentioned just some of them
> > (see my previous posting in this thread)
> > 
> > Aleksey
> > 
> > 
> > 
> > 
> 
> My understanding is that if you do not enable the cache feature
> of the sequence, there will be no gaps.  Of course, disabling the
> cache slows the transaction.

You are extremly right:-) Nevertheless, don't forget of another problem:
you CAN'T call curval() before nextval(). So, every time you want to check
the maximum row number in the table you should increase sequence counter
using nextval() -- here is real source of gaps, and I don't know if it is
possible to overcome it :_( The only idea I have for now is to check 
last_value field of the sequence instead of calling curval()...

But again, what to do with nextval() calls during rollbacked transaction??

Aleksey.




view thread (7+ messages)

Message-ID: <Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com>
Permalink:  ../Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com/
Also on:    postgresql.org/message-id/Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients from your mail client; the link fills in
  the headers below: reply via email
  Some clients, webmail especially, drop In-Reply-To and References from
  a link, and the reply then starts a new thread: use the mbox below.

  To: pgsql-sql@postgresql.org
  Cc: postgres@nest.bistbn.com, marc@fallon.classyad.com, grim@argh.demon.co.uk
  Subject: Re: [SQL] abusing an aggregate funct
  In-Reply-To: <Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com>
  References: <Pine.LNX.4.02A.9811151101140.8582-100000@fallon.classyad.com> <Pine.BSF.4.02.9811161009520.14966-100000@nest.bistbn.com>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox