[OSM-talk] Planet Dump
nickblack1 at gmail.com
Fri Mar 9 09:31:05 GMT 2007
On 3/9/07, Keith Sharp <kms at passback.co.uk> wrote:
> On Fri, 2007-03-09 at 09:08 +0000, Nick Whitelegg wrote:
> > On Friday 09 Mar 2007 08:23, Ray Booysen wrote:
> > > Apologies if this has been brought up before.
> > >
> > > Is it possible to start generating SQL dumps of the same dataset as the
> > > planet.osm file? For me, its easier to download an SQL dump and run
> > > against a server after an export than it would be to parse the XML. I'm
> > > sure we can create it, but not sure on the logistics.
> > >
> > > Any comments?
> > >
> > > Regards
> > > Ray
> > I would very much support this too - specifically the PostGIS dump which I
> > believe is generated weekly anyhow. I've now managed to get the
> > postgis/mapnik setup working successfully on my own server - and the only
> > thing that stops me generating it from planet is the memory requirements of
> > the osm2pgsql utility which my server can't handle :-)
> I was about to try setting up PostGIS + Planet + Mapnik on my home
> system, how much memory does osm2pgsql need?
> I would second Raphaëls suggestion that two dumps be made available: the
> full planet file as at present and a file containing the delta since the
> previous planet dump.
Yep, for sure, i've been thinking this would be cool, but haven't sat
down and thought about how to extract diffs. With SQL, it would be
fine, but how can an XML file just show changes? I guess you would
just run planet.rb on objects in the db that have been modified since
the last time planet.rb was run, and leave it up to the clients to
integrate this into their databases/apps or whatever - unless there's
a more intelligent way to do this that I'm missing...
This of course is more work for the OSM sys admins, so if we just got
a MySQL dump, we could crack on with this sort of stuff ourselves.
> talk mailing list
> talk at openstreetmap.org
More information about the talk