<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso-8859-1">
<META NAME="Generator" CONTENT="MS Exchange Server version 6.5.7652.24">
<TITLE>RE: [postgis-devel] More Cascade Union Adventures</TITLE>
</HEAD>
<BODY>
<!-- Converted from text/plain format -->
<P><FONT SIZE=2>Sure - its just the one on the MassGIS site<BR>
<BR>
<A HREF="ftp://data.massgis.state.ma.us/pub/shape/state/towns.exe">ftp://data.massgis.state.ma.us/pub/shape/state/towns.exe</A><BR>
<BR>
Shucks - I guess that means I don't win a prize huh :(<BR>
<BR>
Thanks,<BR>
Regina<BR>
<BR>
-----Original Message-----<BR>
From: postgis-devel-bounces@postgis.refractions.net on behalf of Martin Davis<BR>
Sent: Wed 8/13/2008 6:40 PM<BR>
To: PostGIS Development Discussion<BR>
Subject: Re: [postgis-devel] More Cascade Union Adventures<BR>
<BR>
I'm pretty sure that OJ is *not* calling CascadeUnion, but is using a<BR>
special-purpose, similar algorithm developed by Michael Michaud. I<BR>
just looked at the OJ src code to confirm, and this looks to be the<BR>
situation. I think he tested his code against the JTS CascadedUnion and<BR>
decided that the JTS algorithm is still faster.<BR>
<BR>
I just tried the OJ Union as well as the JTS CascadedUnion on the<BR>
32278-feature dataset. Results were:<BR>
<BR>
OJ Union: 27 sec<BR>
JTS CascUnion: 17.7 sec<BR>
<BR>
So JTS still wins, but they're not too far apart.<BR>
<BR>
Just for grins, can you send me your Mass towns dataset to try?<BR>
<BR>
Obe, Regina wrote:<BR>
><BR>
> Martin,<BR>
><BR>
> Just loaded that plugin - unless I put it in the wrong folder. - it<BR>
> just seemed to create an option called aggregation-options under<BR>
> plugins which seems to require two layers and does union, count<BR>
> between the 2 layers. So nope doesn't sound like Cascaded union and<BR>
> didn't seem to change anything with the union function.<BR>
><BR>
> If OpenJump union function isn't doing cascaded union, then what is<BR>
> that count down thing for. Doing the 30,000 it displays something<BR>
> like this<BR>
><BR>
> Computing Union<BR>
> 1/8 (32,278)<BR>
> 2/8 (8070)<BR>
> 3/8 ... /2018<BR>
> 86/505<BR>
> 1/127<BR>
> 6/8 (2/32)<BR>
> 8/8 (2/3)<BR>
> 3/3 (8/8)<BR>
><BR>
> I also thought its speed of unioning of Mass towns was pretty impressive.<BR>
><BR>
> Thanks,<BR>
> Regina<BR>
> -----Original Message-----<BR>
> From: postgis-devel-bounces@postgis.refractions.net on behalf of Obe,<BR>
> Regina<BR>
> Sent: Wed 8/13/2008 5:40 PM<BR>
> To: PostGIS Development Discussion<BR>
> Subject: RE: [postgis-devel] More Cascade Union Adventures<BR>
><BR>
> I assumed it was since it was counting down like in some sort of<BR>
> upside down pyramid<BR>
><BR>
> 500<BR>
> 255<BR>
> 10<BR>
> :<BR>
> :<BR>
><BR>
> So it seemed like it was doing some sort of division of the<BR>
> geometries. I'll give the below<BR>
> a try. Maybe I misunderstood what that counting was for.<BR>
><BR>
><BR>
><BR>
> -----Original Message-----<BR>
> From: postgis-devel-bounces@postgis.refractions.net on behalf of<BR>
> Martin Davis<BR>
> Sent: Wed 8/13/2008 4:49 PM<BR>
> To: PostGIS Development Discussion<BR>
> Subject: Re: [postgis-devel] More Cascade Union Adventures<BR>
><BR>
> Yep, I don't Hausdorff distance is available in very many places. Pity,<BR>
> because it is a very useful metric for comparing geometry. The full<BR>
> Hausdorff distance is quite challenging to implement, but I have made a<BR>
> VertexHausdorffDistance approximation which is just as useful in most<BR>
> situations, and is a lot simpler to implement (and faster to run).<BR>
><BR>
> By the way, are you sure that OpenJUMP is using CascadedUnion? AFAIK it<BR>
> didn't in the past... Michael Michaud has just released (today!) an<BR>
> extension which I think does use CascadedUnion - so you might want to<BR>
> try that.<BR>
><BR>
> <A HREF="http://geo.michaelm.free.fr/OpenJUMP/resources/aggregation-0.1.jar">http://geo.michaelm.free.fr/OpenJUMP/resources/aggregation-0.1.jar</A><BR>
><BR>
> When I did the original testing with the 30K polygon dataset that you're<BR>
> using, I was getting times of around 20 sec using CascadedUnion...<BR>
><BR>
><BR>
><BR>
> Obe, Regina wrote:<BR>
> ><BR>
> > Martin,<BR>
> > Pardon my ignorance.<BR>
> ><BR>
> > I don't see a Hausdorff distance in OpenJump or maybe it goes by a<BR>
> > more verbose name and the descriptions I read about Hausdorff<BR>
> > distances are greek to me.<BR>
> ><BR>
> > Well the Mass town test seems to pass my trivial test exercises. It<BR>
> > looks like massachusetts with no visually apparent gaps, has the same<BR>
> > number points in all cases, similar area<BR>
> ><BR>
> > both num points - 476026<BR>
> > both num geometries - 694<BR>
> > area (ST_CascadeUnion - 2.094208570266725E10)<BR>
> > area (JTS - 2.0942085702666965E10)<BR>
> ><BR>
> ><BR>
> > -----Original Message-----<BR>
> > From: postgis-devel-bounces@postgis.refractions.net on behalf of<BR>
> > Martin Davis<BR>
> > Sent: Wed 8/13/2008 11:51 AM<BR>
> > To: PostGIS Development Discussion<BR>
> > Subject: Re: [postgis-devel] More Cascade Union Adventures<BR>
> ><BR>
> > I wouldn't expect to have the results be exactly equal - the union code<BR>
> > is likely to be input-order dependent.<BR>
> ><BR>
> > And the difference in areas doesn't seem surprising either - it's way<BR>
> > down in the small decimal places, which would occur even with slight<BR>
> > differences in the geometry.<BR>
> ><BR>
> > A more revealing test would be to compute the Hausdorff distance between<BR>
> > the union boundaries - that would show if they differed by very much,<BR>
> > and where. PostGIS doesn't have this - I can't remember whether<BR>
> > OpenJUMP does or not. JEQL has this operation, too.<BR>
> ><BR>
> > Obe, Regina wrote:<BR>
> > > Now testing all including JTS 1.9.0 (OpenJump) on a Win XP runing<BR>
> > > PostgreSQL 8.3.1, PostGIS 1.3.3, Geos 3.0.0. It appears using array<BR>
> > > trumps all, Cascade aggregate union is a vast improvement over<BR>
> ST_Union<BR>
> > > (ST_Union I didn't bother testing of course because it owuld never<BR>
> > > finish on this test), but evidentally the array accum calls give a<BR>
> major<BR>
> > > penalty.<BR>
> > ><BR>
> > > I found some things I found a bit possibly disturbing. Maybe its just<BR>
> > > the nature of unioning in different orders. I compared all 3 outputs<BR>
> > > and none of them are binary equal or even ST_Equals for that matter.<BR>
> > ><BR>
> > > However all 3 give same basic stats using OpenJump - e.g<BR>
> > > All result in = 32972 pts <BR>
> > > components = 10<BR>
> > > lengths and areas are off by a bit<BR>
> > > length = agg cascade union = 17.262684721407624, k nested union =<BR>
> > > 17.262684721407688,<BR>
> > > jts = 17.262684721406348<BR>
> > ><BR>
> > > Should I be bothered by any of these or are they just rounding errors?<BR>
> > ><BR>
> > > --Other odd thing is that the array approach is not as good as it<BR>
> was on<BR>
> > > my other machine bu the aggregate union performs better. I'll just<BR>
> > > chuck this off to different postgresql version, and memory settings.<BR>
> > ><BR>
> > > Thanks,<BR>
> > > Regina<BR>
> > ><BR>
> > ><BR>
> > > -- 209,563 | 198,391 ms - SELECT 198391/1000.00/60 = 3.31 minutes<BR>
> > ><BR>
> > > SELECT ST_CascadeUnion(the_geom)<BR>
> > > FROM (SELECT the_geom FROM sample_poly) As foo;<BR>
> > ><BR>
> > > -- 48,594 ms | 47,688 ms = 48 secs<BR>
> > > SELECT st_unitecascade_garray_sort(ARRAY(SELECT the_geom FROM<BR>
> > > sample_poly));<BR>
> > ><BR>
> > > -- 74,515 ms | 74,922 ms = SELECT 74515/1000.00/60 = 1.24 minutes<BR>
> > > SELECT ST_Union(the_geom) AS the_geom, 'nested union'<BR>
> > > FROM (<BR>
> > > SELECT min(id) AS id, ST_Union(the_geom) AS the_geom<BR>
> > > FROM (<BR>
> > > SELECT min(id) AS id, ST_Union(the_geom) AS the_geom<BR>
> > > FROM (<BR>
> > > SELECT min(id) AS id, ST_Union(the_geom) AS the_geom<BR>
> > > FROM (<BR>
> > > SELECT min(id) AS id, ST_Union(the_geom) AS the_geom<BR>
> > > FROM (SELECT the_geom, id FROM sample_poly) As foo<BR>
> > > GROUP BY round(id/10)<BR>
> > > ORDER BY id) AS tmp1<BR>
> > > GROUP BY round(id/100)<BR>
> > > ORDER BY id) AS tmp2<BR>
> > > GROUP BY round(id/1000)<BR>
> > > ORDER BY id) AS tmp3<BR>
> > > GROUP BY round(id/10000)<BR>
> > > ORDER BY id) AS tmp4<BR>
> > > GROUP BY round(id/100000);<BR>
> > ><BR>
> > ><BR>
> > > -- Open Jump JTS 1.9.0 Win XP<BR>
> > > --After loading running union across the whole set<BR>
> > > --1.14 minutes<BR>
> > > --Loading database query<BR>
> > > SELECT ST_AsBinary(the_geom)<BR>
> > > FROM sample_poly;<BR>
> > ><BR>
> > > -----------------------------------------<BR>
> > > The substance of this message, including any attachments, may be<BR>
> > > confidential, legally privileged and/or exempt from disclosure<BR>
> > > pursuant to Massachusetts law. It is intended<BR>
> > > solely for the addressee. If you received this in error, please<BR>
> > > contact the sender and delete the material from any computer.<BR>
> > ><BR>
> > > _______________________________________________<BR>
> > > postgis-devel mailing list<BR>
> > > postgis-devel@postgis.refractions.net<BR>
> > > <A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
> > ><BR>
> > ><BR>
> ><BR>
> > --<BR>
> > Martin Davis<BR>
> > Senior Technical Architect<BR>
> > Refractions Research, Inc.<BR>
> > (250) 383-3022<BR>
> ><BR>
> > _______________________________________________<BR>
> > postgis-devel mailing list<BR>
> > postgis-devel@postgis.refractions.net<BR>
> > <A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
> ><BR>
> ><BR>
> ><BR>
> > ------------------------------------------------------------------------<BR>
> ><BR>
> > _______________________________________________<BR>
> > postgis-devel mailing list<BR>
> > postgis-devel@postgis.refractions.net<BR>
> > <A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
> > <BR>
><BR>
> --<BR>
> Martin Davis<BR>
> Senior Technical Architect<BR>
> Refractions Research, Inc.<BR>
> (250) 383-3022<BR>
><BR>
> _______________________________________________<BR>
> postgis-devel mailing list<BR>
> postgis-devel@postgis.refractions.net<BR>
> <A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
><BR>
><BR>
> ------------------------------------------------------------------------<BR>
><BR>
> _______________________________________________<BR>
> postgis-devel mailing list<BR>
> postgis-devel@postgis.refractions.net<BR>
> <A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
> <BR>
<BR>
--<BR>
Martin Davis<BR>
Senior Technical Architect<BR>
Refractions Research, Inc.<BR>
(250) 383-3022<BR>
<BR>
_______________________________________________<BR>
postgis-devel mailing list<BR>
postgis-devel@postgis.refractions.net<BR>
<A HREF="http://postgis.refractions.net/mailman/listinfo/postgis-devel">http://postgis.refractions.net/mailman/listinfo/postgis-devel</A><BR>
<BR>
</FONT>
</P>
</BODY>
</HTML>