DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=mime-version:in-reply-to:references:date:message-id:subject:from:to
	:cc:content-type:content-transfer-encoding;
	b=YbEPOJ4xW2f/sJTtYZtV/T8FABR/wP0Wn5GrKXwS1q0qIulUGMkp5KS1S27vVdj8a8
	X1lL7QWohPjQrp9bRTDEdAMzgu7a/K/VK3ZKr+cBeaL8ldSSOCBvC8LKFcjCfOzdXH0Z
	rlWAa+0swbogJ5g6LkfrrEMyY0FUTvAFuouvM=
MIME-Version: 1.0
In-Reply-To: <p2k2f4958ff1004240738i91f8b5eco8b4571f994e15254@mail.gmail.com>
References: <n2ybf6923ed1004211316ib844d82bn7a63bd3292ca80a0@mail.gmail.com>
	<n2sb42b73151004220711mc2135db6heccf53ec1fb0136e@mail.gmail.com>
	<q2yb42b73151004220742gbb7ae4adyc3781cedf98e8559@mail.gmail.com>
	<m2gbf6923ed1004221357ua9ab54a3s92dfdbb59cfef832@mail.gmail.com>
	<n2qbf6923ed1004231342l61f766fax8fb2a482aa3b0faf@mail.gmail.com>
	<r2wb42b73151004231401r120510e8ne5ece4cdda768d5a@mail.gmail.com>
	<x2kbf6923ed1004231739qe92d06c2o557d4e31e66ba3d7@mail.gmail.com>
	<v2ybf6923ed1004231931gc02a4368wa00bbc0fdf8d4543@mail.gmail.com>
	<s2lb42b73151004240623h78726052h4f30fade77f8b572@mail.gmail.com>
	<p2k2f4958ff1004240738i91f8b5eco8b4571f994e15254@mail.gmail.com>
Date: Sat, 24 Apr 2010 11:20:12 -0400
Message-ID: <y2kb42b73151004240820u5a0082dt2201eef3c4dafa87@mail.gmail.com>
Subject: Re: Replacing Cursors with Temporary Tables
From: Merlin Moncure <mmoncure@gmail.com>
To: =?ISO-8859-2?Q?Grzegorz_Ja=B6kiewicz?= <gryzman@gmail.com>
Cc: Eliot Gable <egable+pgsql-performance@gmail.com>,
	pgsql-performance@postgresql.org
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

2010/4/24 Grzegorz Ja=C5=9Bkiewicz <gryzman@gmail.com>:
>
>
> On Sat, Apr 24, 2010 at 2:23 PM, Merlin Moncure <mmoncure@gmail.com> wrot=
e:
>>
>> Well, you missed the most important part: not using cursors at all.
>> Instead of declaring a cursor and looping it to build the array, build
>> it with array(). =C2=A0That's what I've been saying: arrays can complete=
ly
>> displace both temp tables _and_ cursors when passing small sets around
>> functions.
>>
> with huge emphasis on the word small.

The rule of thumb I use is 10000 narrow records (scalars, or very
small composites) or 1000 wide/complex records.  I routinely pass
extremely complex (3-4 levels nesting) nested composite arrays to the
client for processing -- it is extremely efficient and clean.  This of
course is going to depend on hardware and other factors.

merlin