From: Steve Brett <steve.brett@e-mis.com>
To: pgsql-sql@postgresql.org
Subject: Re: persistent connections, AOLserver (Was: maybe Offtopic : PostgreSQL & PHP ?)
Date: Fri, 30 Nov 2001 10:23:01 -0000
Message-ID: <9u7mtf$fd2$1@news.tht.net> (raw)
References: <20011119083547.A22031@cc.usu.edu>
is there a limit/guide to the number of inserts you should use in a
transaction block ?
i have an app at the minute written indelphi that moves data from a cache
database and a sql server databse and produces one definitive customer
management database for the company i work for.
the transfer is quite slow but i do use single inserts (not through a lack
of knowledge for what a database is for though :-) and maybe wrapping in a
transaction block is the answer.
i'm moving (parsing and matching) pretty small recorsets (10,000 in one and
approx 60,000 in the other).
Steve
"Roberto Mello" <rmello@cc.usu.edu> wrote in message
news:20011119083547.A22031@cc.usu.edu...
> On Wed, Apr 18, 2001 at 08:28:14PM +0200, Mathijs Brands wrote:
> >
> > I've seen both MySQL and PostgreSQL give up (MySQL just hung, pgsql
> > cored) when I tried simple selects from a couple of hundred concurrent
> > connections; no transactions or other fancy stuff there. I think I was
> > using MySQL 3.22.?? and pgsql 6.5.3, so more modern versions may well be
> > able to cope with these numbers of connections. It's been more than a
> > year since I last tried it.
>
> PG 7.0 was released almost 2 years ago, so you had a very old version when
> you tried.
>
> > I don't know what the current state of affairs is, but it is my
> > understanding that, while pgsql performs admirably on tasks which mostly
> > read data, pgsql isn't really able to cope (performance wise) with an
> > application that has a very high insert to select ratio, such as OLTP.
>
> This is simply not true. It may have been true in the past, but not since
> PG 7.0 came out.
>
> A mistake that I see MySQL users do frequently is that when they have a
> bulk of inserts to do, they don't use the COPY command to bulk load the
> data and/or they don't wrap the inserts into one (or several) transaction
> blocks, leaving each INSERT in its own transaction, which obviously will
> hurt performance.
>
> They do that because they usually have no idea of what a database is for,
> besides being an SQL interface to their file systems. Exactly what MySQL
> was until very recently (and thanks to the work of third-parties, because
> the MySQL team itself couldn't care less about real features).
>
> -Roberto
>
> --
> +----| http://fslc.usu.edu USU Free Software & GNU/Linux Club |------+
> Roberto Mello - Computer Science, USU - http://www.brasileiro.net
> http://www.sdl.usu.edu - Space Dynamics Lab, Developer
> TAFB -> Text Above Fullquote Below
>
> ---------------------------(end of broadcast)---------------------------
> TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
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: steve.brett@e-mis.com
Subject: Re: persistent connections, AOLserver (Was: maybe Offtopic : PostgreSQL & PHP ?)
In-Reply-To: <9u7mtf$fd2$1@news.tht.net>
* 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