Received: from localhost (unknown [200.46.204.183])
	by mail.postgresql.org (Postfix) with ESMTP id 9CC42633F00
	for <pgsql-performance-postgresql.org@mail.postgresql.org>;
	Tue,  3 Mar 2009 08:12:50 -0400 (AST)
Received: from mail.postgresql.org ([200.46.204.86])
	by localhost (mx1.hub.org [200.46.204.183]) (amavisd-maia, port 10024)
	with ESMTP id 99551-04
	for <pgsql-performance-postgresql.org@mail.postgresql.org>;
	Tue,  3 Mar 2009 08:12:46 -0400 (AST)
X-Greylist: from auto-whitelisted by SQLgrey-1.7.6
Received: from broadway.cambrient.com (broadway.cambrient.com
	[196.211.141.188])
	by mail.postgresql.org (Postfix) with ESMTP id BB23C633EAB
	for <pgsql-performance@postgresql.org>;
	Tue,  3 Mar 2009 08:12:45 -0400 (AST)
Received: from 192.168.0.69 ([192.168.0.69])
	by broadway.cambrient.com (JAMES SMTP Server 2.3.1) with SMTP ID 812
	for <pgsql-performance@postgresql.org>;
	Tue, 3 Mar 2009 14:18:27 +0200 (SAST)
Message-ID: <49AD1EBD.1050105@cambrient.com>
Date: Tue, 03 Mar 2009 14:12:45 +0200
From: Hans Liebenberg <hans@cambrient.com>
Reply-To: hans@cambrient.com
User-Agent: Thunderbird 2.0.0.19 (Windows/20081209)
MIME-Version: 1.0
To: pgsql-performance@postgresql.org
Subject: Substring search using "exists" with a space in the search term
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: Maia Mailguard 1.0.1
X-Spam-Status: No, hits=1.673 tagged_above=0 required=5
	tests=HTML_MESSAGE=0.001, MIME_HTML_ONLY=1.672
X-Spam-Level: *
X-Archive-Number: 200903/23
X-Sequence-Number: 32945

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
</head>
<body bgcolor="#ffffff" text="#000000">
<span id="intelliTxt">Hi,<br>
<br>
I have come across a weird bug (i think) in postgres 8.1.11 (possibly
others)<br>
<br>
Without going into my table structure detail I will demonstrate the
problem by showing the select statements:<br>
<br>
The following statement:<br>
SELECT count(*)<br>
FROM object o, object_version v, object_type ot <br>
where v.id = o.active_versionid and ot.id = o.object_typeid and
o.is_active ='t' and (o.is_archived = 'f' or o.is_archived is null) <br>
and o.is_published = 't' and ot.object_type_typeid &lt;&gt; 1 <br>
<br>
and exists (<br>
select ova.object_versionid from attribute_value av,
object_version_attribute ova where ova.attribute_valueid=av.id and
object_versionid = v.id <br>
and (upper(av.text_val) like <b>'%KIWI%'</b>) )<br>
<br>
<br>
runs fine and executes with success.<br>
BUT now this is the strange bit, if I have a space in my search term
then postgres hangs for an indefinite period: eg:<br>
<br>
SELECT count(*)<br>
FROM object o, object_version v, object_type ot <br>
where v.id = o.active_versionid and ot.id = o.object_typeid and
o.is_active ='t' and (o.is_archived = 'f' or o.is_archived is null) <br>
and o.is_published = 't' and ot.object_type_typeid &lt;&gt; 1 <br>
<br>
and exists (<br>
select ova.object_versionid from attribute_value av,
object_version_attribute ova where ova.attribute_valueid=av.id and
object_versionid = v.id <br>
and (upper(av.text_val) like <b>'%KIWI FRUIT%'</b>) )<br>
<br>
<br>
Yet, if I modify the "exists" to an "in" all works well , as follows<br>
<br>
SELECT count(*)<br>
FROM object o, object_version v, object_type ot <br>
where v.id = o.active_versionid and ot.id = o.object_typeid and
o.is_active ='t' and (o.is_archived = 'f' or o.is_archived is null) <br>
and o.is_published = 't' and ot.object_type_typeid &lt;&gt; 1 <br>
<br>
and v.id in (<br>
select ova.object_versionid from attribute_value av,
object_version_attribute ova where ova.attribute_valueid=av.id <br>
and (upper(av.text_val) like <b>'%KIWI FRUIT%'</b>) )<br>
<br>
<br>
So my question is why would a space character cause postgres to hang
when using the exists clause????<br>
<br>
I have tested this on several different servers and mostly get the same
result (v8.08 and v8.1.11) , when I check the execution plan for either
query (space or no space) they are identical.<br>
<br>
An upgrade to 8.3 fixes this, but I am still curious as to what could
cause such bizarre behavior.<br>
<br>
Thanks<br>
Hans</span>
</body>
</html>