[postgis-devel] GeomUnion Performance Info

strk at refractions.net strk at refractions.net
Sat Jun 25 00:40:03 PDT 2005


On Fri, Jun 24, 2005 at 03:38:56PM -0400, Bill Binko wrote:
> On Thu, 23 Jun 2005 strk at refractions.net wrote:
> 
> > > It will take at least tomorrow to do that.  However, I'll aim for being 
> > > complete before the weekend in case there's development to be done then.
> 
> I'm sorry strk: my day job (you know the one with the pay check) has 
> swamped me.  I will try to get to this this evening, but it might be 
> tomorrow.
> 
> Again: sorry!

No problem. I'm attacking the problem from another prospective
and Martin Davis suggested a way which actually runs about 3 times
faster.

For everyone interested, please compare these calls, which should
return the same result:

	memgeomunion(the_geom);
	geomunion(the_geom);
	buffer(collect(the_geom), 0);

Last version should be faster and not be affected by input order.
Graph is built once!

I think we can have geomunion() internally use the buffer(0) method,
which could be faster then the explicit buffer(collect()) call
in that should require less copies of data.

I'm experimenting this.

--strk;




More information about the postgis-devel mailing list