pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: William Scott Jordan <wsjordan@brownpapertickets.com>
To: pgsql-sql@postgresql.org
Subject: Referencing tables are grinding UPDATE to a halt
Date: Mon, 15 Jan 2007 13:26:04 -0800
Message-ID: <7.0.1.0.2.20070115124931.06433838@pandimensional.org> (raw)

Hey all!

I'm having some trouble with a simple update on a table that only has 
about 250,000 rows in it.  The table itself looks something like:

CREATE TABLE price_details (
price_detail_id int PRIMARY KEY,
private bool
) ;

There is one table that references price_details, but isn't affected 
by the "private" column, and one table that references this second 
table.  They look something like:

CREATE TABLE prices (
p_id int PRIMARY KEY,
price_detail_id int NOT NULL REFERENCES price_details ON DELETE CASCADE,
max_sales int
) ;

CREATE INDEX prices_price_detail_id  ON prices (price_detail_id) ;

CREATE TABLE sales (
sales_id int PRIMARY KEY,
p_id int NOT NULL REFERENCES prices ON DELETE CASCADE,
sales int
) ;

CREATE INDEX sales_p_id ON sales (p_id) ;

I'm trying to do a simple update to the "private" column in the 
price_details table, which I expected to take a few seconds at 
most.  After 10 minutes, I gave up and ran explain, with this as the result:


EXPLAIN UPDATE price_details SET private = 't' WHERE private = 'f' ;
----------------------------------------------------------------------------------------
  Nested Loop  (cost=2663.45..363527947.70 rows=118759098 width=50)
    Join Filter: (subplan)
    ->  Seq Scan on sales  (cost=0.00..3685.27 rows=54627 width=42)
    ->  Materialize  (cost=2663.45..2706.93 rows=4348 width=12)
          ->  Seq Scan on price_details  (cost=0.00..2663.45 
rows=4348 width=12)
                Filter: (private  = false)
    SubPlan
      ->  Index Scan using prices_price_detail_id on 
prices  (cost=0.00..3.01 rows=1 width=4)
            Index Cond: (price_detail_id = $0)

  Seq Scan on price_details  (cost=0.00..2663.45 rows=4348 width=508)
    Filter: (private = false)
(12 rows)
----------------------------------------------------------------------------------------

So it looks to me like the postgres is checking this table against 
the table that references it, and the table that reference that 
table, making what should be a very easy transaction into something 
unusable.  Is there any way to avoid this without losing proper referencing?

Any suggestions would be appreciated.

-Scott




Message-ID: <7.0.1.0.2.20070115124931.06433838@pandimensional.org>
Permalink:  ../7.0.1.0.2.20070115124931.06433838@pandimensional.org/
Also on:    postgresql.org/message-id/7.0.1.0.2.20070115124931.06433838@pandimensional.org

 · 

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-sql@postgresql.org
  Cc: wsjordan@brownpapertickets.com
  Subject: Re: Referencing tables are grinding UPDATE to a halt
  In-Reply-To: <7.0.1.0.2.20070115124931.06433838@pandimensional.org>

* 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