[GRASS5] r.random and topology: building areas etc?

[Markus Neteler]

Hi,

the problem below is still unsolved... do we need to build full topology
for points?

Related to that: Exporting points with v.our.ogr is very slow,
do we have/need to integrate spatial index for points?
Example: Exporting 120k points takes 2 minutes on a 2.5GHz machine
to reach only 1 percent...

grass61 ~/grassdata/spearfish60/user1/
g.region rast=elevation.10m -p
r.random elevation.10m vect=random120k n=120000
time v.out.ogr random120k dsn=random120k.shpolay=random120k type=point
Exporting 120000 points/lines...
real    233m18.811s
user    230m38.497s
sys     1m41.950s

Line export is really fast compared to points export.

Any ideas?

Markus

On Thu, Jun 23, 2005 at 04:50:30AM +0200, Markus Neteler wrote:
> Hi,
> 
> I was generating 3.x mio points to test v.in.ascii and discovered,
> that r.random build full topology which is just a waste of time.
> How to change that?
> 
> r.random elevation.10m  n=3200000 vector_output=random3_2mio
> Collecting Stats ...  100%
> Writing site file [random3_2mio] ...  100%
> Erzeuge Topologie ...
> 3200000 primitives registriert
> Building areas:    7%
> ...
> 
> Should we use Vect_build_partial()?
> 
> Markus
>