Re: [postgis-devel] JDBC & WKB - 400% overhead

From: Markus Schaber <schabios(at)logi-track(dot)com>
To: Kris Jurka <books(at)ejurka(dot)com>
Cc: PostGIS Development Discussion <postgis-devel(at)postgis(dot)refractions(dot)net>, "pgsql-jdbc(at)postgresql(dot)org" <pgsql-jdbc(at)postgresql(dot)org>, David Blasby <dblasby(at)openplans(dot)org>
Subject: Re: [postgis-devel] JDBC & WKB - 400% overhead
Date: 2005-02-24 09:39:14
Message-ID: 421DA0C2.10407@logi-track.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-jdbc

Hi, Kris,

Kris Jurka schrieb:

>>Maybe your postgresql driver only uses text transfer up to now.
>>asBinary() returns a byte array, so this should be transferred more
>>efficiently.
>
> The driver receives all data as text. To be able to receive binary data
> we must support binary data for all types because we do not know what type
> we will be getting back from the SELECT before it is run. Getting that
> information would require an extra network round trip so we don't want to
> do that. Doing all binary transfer is on the 8.1 todo list, but I've yet
> to really get started on it.

Oh, I thought that binary transmit was implemented in 8.0 :-(

Then, currently, the better way to receive geometries may be to use the
new canonical text rep for geometries (hex-encoded extended WKB) which
is only two times as large compared to pure binary rep.

As there are at least plans to use caching of prepared statements, it
also should be possible to cache this information together with the
prepared statement. This would reduce the number of roundtrips.

Markus

--
markus schaber | dipl. informatiker
logi-track ag | rennweg 14-16 | ch 8001 zürich
phone +41-43-888 62 52 | fax +41-43-888 62 53
mailto:schabios(at)logi-track(dot)com | www.logi-track.com

In response to

Responses

Browse pgsql-jdbc by date

  From Date Subject
Next Message Kris Jurka 2005-02-24 10:02:34 Re: [postgis-devel] JDBC & WKB - 400% overhead
Previous Message Oliver Jowett 2005-02-24 05:12:02 Re: Socket timeouts (probably)