[postgis-users] Storage efficiency of point and line data

[Michael Graff]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

David Blasby <dblasby at refractions.net> writes:

> Yes - as you can see from the actual structure, there is quite a bit of
> overhead per-geometry.  The 24 bytes per points is because each point is
> being stored as 3 double coordinates {x,y,z}.  A DOUBLE is 8 bytes, so
> each point takes 24 bytes.
> 
> I dont see how you can get 6 bytes per points - thats only 2 bytes per
> ordinate.  You will not have much precision.

I know my data, so I don't need to generalize to everything.  This is
the TIGER data, so each point actually comes out of those files as a
value that fits in a 32-bit integer.  All I need to do is divide it
down (or multiply up the query) by 1 million.

> > Is the storage format fairly efficient, and I'm simply storing a whole
> > lot of data?
> 
> 40,000,000 rows is a LOT of data!

It's actually closer to 47 million for the complete path data, and
I've not started on the polygon data yet.  I'm trying hard to fit this
all onto a 40 G laptop drive, so the GPS is more mobile than a 3U box.
It's challenging.

- --Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (NetBSD)
Comment: See http://www.flame.org/~explorer/pgp for my keys

iD8DBQE9x/lFl6Nz7kJWYWYRArS8AJ0bWsijo/apNOLlXxSjMG17DURp1gCdFkyG
4oV7QcfEtk6GqIVg8xTgPHk=
=JMGO
-----END PGP SIGNATURE-----