pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Hartmut Raschick <hartmut.raschick@ke-elektronik.de>
Cc: pgsql-performance@postgresql.org, Rod Taylor <pg@rbt.ca>
Subject: Re: TRUNCATE veeeery slow compared to DELETE in 7.4
Date: Mon, 15 Dec 2003 13:21:47 -0500
Message-ID: <13898.1071512507@sss.pgh.pa.us> (raw)
In-Reply-To: <3FD99CDA.8161D80A@ke-elektronik.de>
References: <3FD484B6.EF39A92F@ke-elektronik.de>
	<200312100918.02501.josh@agliodbs.com>
	<8000.1071086054@sss.pgh.pa.us>
	<3FD99CDA.8161D80A@ke-elektronik.de>

Hartmut Raschick <hartmut.raschick@ke-elektronik.de> writes:
> [ TRUNCATE is much slower in 7.4 than in 7.3 ]

After looking into this, I think this is because when Rod Taylor
reimplemented TRUNCATE to make it transaction-safe, he essentially
turned it into a variant of CLUSTER.  It is slow because it is creating
and deleting dummy tables and indexes.  I think this is not really
necessary and it could be done better while still being
transaction-safe.  All we really need is to create a new empty table
file, update the table's pg_class row with the new relfilenode, mark
the old file for deletion, and then run REINDEX TABLE (which will
perform similar shenanigans with the indexes).

			regards, tom lane



view thread (5+ messages)

Message-ID: <13898.1071512507@sss.pgh.pa.us>
Permalink:  ../13898.1071512507@sss.pgh.pa.us/
Also on:    postgresql.org/message-id/13898.1071512507@sss.pgh.pa.us

 · 

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 using the --to and --cc options:
  reply via email

  To: pgsql-performance@postgresql.org
  Cc: tgl@sss.pgh.pa.us, hartmut.raschick@ke-elektronik.de, pg@rbt.ca
  Subject: Re: TRUNCATE veeeery slow compared to DELETE in 7.4
  In-Reply-To: <13898.1071512507@sss.pgh.pa.us>

* 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