pg.ddx.io  pgsql-general@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Andrus <kobruleht2@hot.ee>
To: pgsql-general <pgsql-general@postgresql.org>
Subject: How to speed up product code and subcode match
Date: Tue, 23 May 2023 09:53:02 +0300
Message-ID: <7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee> (raw)

Hi!

Price list of main products vordlusajuhinnak contains 3 prices for 
product (column toode) and has 39433 products:

     create table vordlusajuhinnak( toode varchar(60), n2 numeric(8,2), 
n3 numeric(8,2), n4 numeric(8,2) );

toode column in unique, may be primary key in table and contains upper 
case letters, digits and - characters.

product table (toode) contains 733021 products:

     CREATE TABLE toode (
         grupp character(1),
         toode character(60) primary key,
         ... lot of other columns
       );

Both tables have pattern indexes to speed up queries:

     CREATE INDEX toode_toode_pattern_idx
         ON toode (toode bpchar_pattern_ops ASC NULLS LAST);
     -- This index is probably not used, should removed:
     CREATE INDEX vordlusajuhinnak_toode_pattern_idx ON 
vordlusajuhinnak(toode bpchar_pattern_ops);

Product table as both main products and subproducts with sizes. Size is 
separated by main product code always by / character:


     SHOE1-BLACK
     SHOE1-BLACK/38
     SHOE1-BLACK/41
     SHOE1-BLACK/42
     SHOE1-BLACK/43
     SHOE2/XXL
     SHOE2/L

Product codes contain upper characers only in this table also.

Trying to get prices for all products using

     create table peatoode as
     select toode.toode , n2, n3, n4
     from toode, vordlusajuhinnak
     where  toode.toode between vordlusajuhinnak.toode and 
vordlusajuhinnak.toode||'/z'

Takes 4.65 hours. How to speed this query up?
Output from explain:

     > "Nested Loop  (cost=0.55..272273178.12 rows=3202240012 width=78)" "
     > ->  Seq Scan on vordlusajuhinnak  (cost=0.00..721.33 rows=39433 
width=32)" "  ->  Index Only Scan using toode_pkey on toode
     > (cost=0.55..6092.62 rows=81207 width=60)" "        Index Cond: (toode
     > >= (vordlusajuhinnak.toode)::bpchar)" "        Filter: 
((toode)::text <= ((vordlusajuhinnak.toode)::text || '/z'::text))"

Using  PostgreSQL 13.2, compiled by Visual C++ build 1900, 64-bit in 
Windows server and psqlODBC driver.
Upgrading Postgres is possible, if this helps.

Tried also using like:

     WHERE toode.toode=vordlusajuhinnak.toode OR
       toode.toode LIKE vordlusajuhinnak.toode||'/%'

Posted also in

https://stackoverflow.com/questions/76311957/how-to-match-character-columns-with-separated-subcolumn...

Andrus.






view thread (5+ messages)  latest in thread

Message-ID: <7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee>
Permalink:  ../7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee/
Also on:    postgresql.org/message-id/7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee

 · 

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-general@postgresql.org
  Cc: kobruleht2@hot.ee
  Subject: Re: How to speed up product code and subcode match
  In-Reply-To: <7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee>
  References: <7181fbab-1b51-c005-2576-d287fbcd784c@hot.ee>

* 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