[Openstreetmap] How do we handle large amount of GPS points?

[Petter Reinholdtsen]

So, I managed to get in touch with one linux user in Oslo, which has
been driving around in the area with a GPS and a laptop for quite some
time now.  He got >7 million GPS coordinates, which he uses to create
a bitmap map of Oslo.  He was very positive to the idea of sharing
this data set with us, and I got a copy of his current database.

Until now, I've been able to collect 40k points myself.  That has been
no problem.  The GPX files are of reasonable size, and qgis could
handle the data just fine.  I'm not equally sure about the 7000k
points I got now.

I guess my question is really, how should we handle such wast amount
of points?  Steve, do you want 7 million points from Norway in your
database?  Ready to handle it?  How do you want it?

I got it as a mysql database dump, with this table definition:

CREATE TABLE track (
  ID bigint(8) NOT NULL auto_increment,
  time datetime NOT NULL default '0000-00-00 00:00:00',
  lat decimal(8,6) NOT NULL default '0.000000',
  lon decimal(8,6) NOT NULL default '0.000000',
  alt int(4) NOT NULL default '0',
  fix smallint(1) NOT NULL default '0',
  speed decimal(5,2) NOT NULL default '0.00',
  trackfile varchar(150) NOT NULL default '',
  PRIMARY KEY  (ID),
  UNIQUE KEY trackfile_2 (trackfile,lat,lon)
) TYPE=MyISAM;

fix is 2 for 2D (3 satellites) and 3 for 3D (>=4 satellites).  I'll
import it into PostgreSQL and do some processing there.