Date: Fri, 2 Dec 2005 16:19:25 -0500
From: Stephen Frost <sfrost@snowman.net>
To: Luke Lonergan <llonergan@greenplum.com>
Cc: David Lang <dlang@invendra.net>, Steve Oualline <soualline@stbernard.com>,
	pgsql-performance@postgresql.org
Subject: Re: Database restore speed
Message-ID: <20051202211925.GF6026@ns.snowman.net>
Mail-Followup-To: Luke Lonergan <llonergan@greenplum.com>,
	David Lang <dlang@invendra.net>,
	Steve Oualline <soualline@stbernard.com>,
	pgsql-performance@postgresql.org
References: <20051202201847.GE6026@ns.snowman.net>
	<BFB5F308.14E82%llonergan@greenplum.com>
Mime-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha1;
	protocol="application/pgp-signature"; boundary="/UZYnPkX7MOZI66P"
Content-Disposition: inline
In-Reply-To: <BFB5F308.14E82%llonergan@greenplum.com>
User-Agent: Mutt/1.5.9i


--/UZYnPkX7MOZI66P
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

* Luke Lonergan (llonergan@greenplum.com) wrote:
> On 12/2/05 12:18 PM, "Stephen Frost" <sfrost@snowman.net> wrote:
> > Just a thought, but couldn't psql be made to use the binary mode of
> > libpq and do at least some of the conversion on the client side?  Or
> > does binary mode not work with copy (that wouldn't suprise me, but
> > perhaps copy could be made to support it)?
>=20
> Yes - I think this idea is implicit in what David suggested, and my respo=
nse
> as well.  The problem is that the way the client does conversions can
> potentially differ from the way the backend does.  Some of the types in
> Postgres are machine intrinsic and the encoding conversions use on-machine
> libraries, each of which preclude the use of client conversion methods
> (without a lot of restructuring).  We'd tackled this problem in the past =
and
> concluded that the parse / convert stage really belongs in the backend.

I've used the binary mode stuff before, sure, Postgres may have to
convert some things but I have a hard time believing it'd be more
expensive to do a network_encoding -> host_encoding (or toasting, or
whatever) than to do the ascii -> binary change.

	Thanks,

		Stephen

--/UZYnPkX7MOZI66P
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (GNU/Linux)

iD8DBQFDkLpdrzgMPqB3kigRAhQNAKCPzO2l/MJNlrzXtlsW+vYU6Y7ZGACePIuF
50XkDsbVloZ+JQgFHOrIbLI=
=Hf7Z
-----END PGP SIGNATURE-----

--/UZYnPkX7MOZI66P--