gmane.comp.gis.postgis

problems cleaning geometry - other options?

William Kyngesburye — 2008-12-04T00:08:45

I'm trying to fix some polygons-with-holes imported from shapefiles where the main polygon is not the first in the list, one of the hole polygons is, making them invalid. These are 3D multipolygons - imported from the SRTM SWBD shapefiles (with GDAL). (file w162n55n if anyone wants to try) The cleanGeometry funtion at: http://postgis.refractions.net/support/wiki/index.php?CleanPolygons isn't working, I get an error: =>select astext(cleangeometry(wkb_geometry)) from swbd where cell = 145018 and wb = 353; NOTICE: Hole lies outside shell at or near point -161.201 55.9992 0 CONTEXT: PL/pgSQL function "cleangeometry" line 15 at IF ERROR: Dimensions mismatch in lwcurve CONTEXT: PL/pgSQL function "cleangeometry" line 19 at assignment I tried forcing the geometry to 2D (st_force_2d()) and I get a different error: NOTICE: Hole lies outside shell at or near point -161.201 55.9992 CONTEXT: PL/pgSQL function "cleangeometry" line 15 at IF ERROR: Exception in LWGEOM2GEOS: curved geometry not suppor

Re: Re: Problem with MultiLineString

2008-12-03T23:38:23

- The reason is documented. The overhead for validating all geometries as entered is generally unecessary. For cases where it is appropriate, an insert trigger using isvalid() can be used to provide this capability. Cheers, BRnet Wood

Re: Problem with MultiLineString

Ken Southerland — 2008-12-03T21:17:52

Thank you! That's got to be it. I knew I just needed a third eye, didn't even see that. A followup question might be why is this geometry allowed to be created if it is then considered invalid by certain functions? Ken Michael Smedberg said: >It looks to me like maybe > >(491215.9577 532660.8051,491215.9577 532660.8051) > >is an invalid line because both points are identical, while > >(491215.957 532660.8051,491215.9577 532660.8051) > >is valid because the points aren't identical? Just a guess...

RE: Problem with MultiLineString

Michael Smedberg — 2008-12-03T20:46:31

It looks to me like maybe (491215.9577 532660.8051,491215.9577 532660.8051) is an invalid line because both points are identical, while (491215.957 532660.8051,491215.9577 532660.8051) is valid because the points aren't identical? Just a guess... -----Original Message----- From: postgis-users-bounces< at >postgis.refractions.net [mailto:postgis-users-bounces< at >postgis.refractions.net] On Behalf Of Ken Southerland Sent: Wednesday, December 03, 2008 12:22 PM To: postgis-users< at >postgis.refractions.net Subject: [postgis-users] Problem with MultiLineString This problem occurs on versions... POSTGIS="1.3.1" GEOS="3.0.0rc4-CAPI-1.3.3" PROJ="Rel. 4.5.0, 22 Oct 2006" POSTGIS="1.3.2" GEOS="3.0.0-CAPI-1.4.1" PROJ="Rel. 4.6.0, 21 Dec 2007" The first query gives... select st_isvalid( st_GeomFromText('MULTILINESTRING((491181.9816 532686.6037,491209.6059 532664.6521),(491215.9577 532660.8051,491215.9577 532660.8051))')); NOTICE: IllegalArgumentException: Number of points must be 0 or >3 st_isvalid ------------ f

Re: Problem with MultiLineString

Ken Southerland — 2008-12-03T20:39:53

Sorry, I copied the wrong error message after the query, that was from a multipolygon issue that I was simultaneously investigating. The real message for this is ... NOTICE: Too few points in geometry component at or near point 491216 532661 Sorry for the misinformation. Still in need of help! :) Ken Ken Southerland wrote:

Problem with MultiLineString

Ken Southerland — 2008-12-03T20:22:17

This problem occurs on versions... POSTGIS="1.3.1" GEOS="3.0.0rc4-CAPI-1.3.3" PROJ="Rel. 4.5.0, 22 Oct 2006" POSTGIS="1.3.2" GEOS="3.0.0-CAPI-1.4.1" PROJ="Rel. 4.6.0, 21 Dec 2007" The first query gives... select st_isvalid( st_GeomFromText('MULTILINESTRING((491181.9816 532686.6037,491209.6059 532664.6521),(491215.9577 532660.8051,491215.9577 532660.8051))')); NOTICE: IllegalArgumentException: Number of points must be 0 or >3 st_isvalid ------------ f ... while the second gives... select st_isvalid( st_GeomFromText('MULTILINESTRING((491181.9816 532686.6037,491209.6059 532664.6521),(491215.957 532660.8051,491215.9577 532660.8051))')); st_isvalid ------------ t ...and the only difference between the two is that the first point of the second line has an x value of 491215.9577 in the first query and 491215.957 in the second. In other words, the last seven is just stripped and it works. What gives here?! Any ideas? Ken Southerland

Re: Get concave hull from an array of longitude inlatitude points

Aleš Kecman — 2008-12-03T18:36:20

longitude;latitude 14.4748633596;46.0878254619 14.4763518858;46.08903795 14.4818432078;46.0881362379 14.4820886727;46.0867787345 14.4819582754;46.0868951105 14.4829486102;46.0874664345 14.4825109958;46.0872485151 14.482238313;46.0859876362 14.4867016967;46.082948631 14.4880592792;46.0829547196 14.4811801222;46.0885201205 14.4814057611;46.0879003201 14.480595963;46.0887603988 14.4796438292;46.0882522106 14.4800623417;46.0877502577 14.4809601134;46.0871605059 14.4866553199;46.082372585 14.4861897845;46.0823794907 14.4859485352;46.0832961472 14.4851890461;46.0829328304 14.4846191494;46.0830382317 14.484966568;46.083219747 14.484826973;46.0843258057 14.4841620095;46.08353103 14.4854708637;46.0846076273 14.4841742002;46.0849976711 14.4836219952;46.084590292 14.484097376;46.0849163473 14.4845219648;46.0851431994 14.4844437989;46.0852058291 14.4843786465;46.0852595199 14.4840787429;46.0855280904 14.4826033827;46.0856653808 14.4834094892;46.0865777702 --------------------------

Re: Get concave hull from an array of longitude inlatitude points

Martin Davis — 2008-12-03T17:46:26

You might also check out this paper: http://www.foss4g.org/index.php/foss4g/2008/paper/view/226/157 It uses a "connect-buffer-simplify" approach. Martin Davis wrote:

Re: Get concave hull from an array of longitude inlatitude points

Martin Davis — 2008-12-03T17:09:05

I think the ideas for the solution have already been posted on this thread - you either need to use alpha shapes or one of the concave hull algorithms. It sounded like pgrouting has an alpha shapes implementation - that would probably be the easiest way forward. Implementing the other algorithms would be non-trivial. It might also help if you posted a picture of what you think the desired output should be, for the sample data you posted. Aleš Kecman wrote:

Re: Get concave hull from an array of longitude inlatitude points

Aleš Kecman — 2008-12-03T11:02:06

Subarea_name;longitude;latitude "Bled";14.0654858532;46.3471854704 "Bled";14.06523413;46.3474713196 "Bled";14.0648965337;46.347459562 "Bled";14.0648286668;46.3476299516 "Bled";14.065647676;46.3476006589 "Bled";14.0650490436;46.3476587451 "Bled";14.0664993488;46.3487052597 "Bled";14.0660859457;46.3485669276 "Bled";14.0656344927;46.3483743005 "Bled";14.065892772;46.3484663817 "Bled";14.0649190402;46.3491871864 "Bled";14.0643206968;46.349227277 "Bled";14.0644745909;46.3493454984 "Bled";14.0668079097;46.3488967172 "Bled";14.0663594854;46.3492889282 "Bled";14.0649514158;46.3495743258 "Bled";14.0659731326;46.349087837 "Bled";14.0652944371;46.3485064774 "Bled";14.0656042191;46.3486259715 "Bled";14.0658881743;46.3487362567 "Bled";14.0662242151;46.3496027215 "Bled";14.0661603657;46.3487744683 "Bled";14.0664025936;46.3490463589 "Bled";14.0668390737;46.3493558228 "Bled";14.067040212;46.3489885837 "Bled";14.0670568563;46.3495375416 "Bled";14.0665596808;46.3497404169 "Bled";14.066727644;46.3

Re: Exploding Polygons

Carl Anderson — 2008-12-03T01:45:01

Have you considered subtracting a buffer around the boundary of each polygon from itself. Kind of similar to st_buffer with a negative distance but a good bit more stable. st_difference(geom,st_buffer(st_boundary(geom),100))) from mypolytable ; st_buffer(geom,-X) will work until X gets large enough to start to get collapsing geometries and then you will get ERROR: getPoint2d_p: point offset out of range C. Randall, Eric wrote:

Re: Exploding Polygons

Stephen Woodbridge — 2008-12-02T22:43:27

I'm not sure if it will work, but have you tried doing a buffer with a negative distance. select st_buffer(the_geom, -100) from table; where -100 is whatever distance you want the edge to be retracted. This would be in the units of the table. -Steve

Exploding Polygons

Randall, Eric — 2008-12-02T22:27:31

Hi all, I'm trying to figure out how to explode a set of polygons to display them with space between. In my case they are municipal boundary polys but could be anything. So far the only approach I can think of is using translate such that: select st_translate(st_translate(poly, -x(st_centroid(poly)), -y(st_centroid(poly))),x(p2) + (x(p2)- x(p1)), y(p2) + (y(p2) - y(p1))) from (select st_centroid(st_collect(geom))as p1 from mypolytable) as s1, (select st_centroid(geom) as p2, geom as poly from mypolytable) as s2 The problem with this is that small polys (Boroughs, etc) relative larger ones often don't get their own space until a large enough multiplier is applied, i.e. x(p2) + (x(p2)- x(p1))*5, y(p2) + (y(p2) - y(p1))*5 , but then there is too much space around the larger polys. Anyone have a suggestion or done this another way? Thanks. think i'm gonna splode..., Eric

R: Postgres/PostGIS and date field

P.Rizzi Ag.Mobilità Ambiente — 2008-12-02T11:21:26

Maybe: select the_geom .......... extract(day from (end_date - now()) >= 10) as green and extract(day from (end_date - now()) < 0) as red .... from .... Bye Paolo Rizzi

Re: Postgres/PostGIS and date field

Mark Cave-Ayland — 2008-12-02T11:16:34

Hi Vinc, This sounds like the casting changes in 8.3 are hitting you here. Have you tried something like this: select the_geom .......... (end_date - now() >= '10 days'::interval) as green and (end_date - now() < '0'::interval) as red .... from .... I'm not sure exactly what units you were using, but you should be able to work with the example above to get what you want. HTH, Mark.

Postgres/PostGIS and date field

Vincenzo Patruno — 2008-12-02T11:04:16

Hi list! In my file .map I wrote something like select the_geom .......... (end_date - now() >= 10) as green and (end_date - now() < 0) as red .... from .... The idea is using different coloured markers for expired points or valid points After upgrading Postgres/Gis with last available versions, the query above doesn't work anymore because I cannot compare date types with integers ....... Any idea? Thanks in advance Vinc

RE: Arc catalog (Arc GIS) import

Sufficool, Stanley — 2008-12-01T18:34:22

Before I get too far on this and run into a roadblock (legally or technically), has anyone tried the Personal Geodatabase linked to a PostgreSQL view? My general plan is to create a view using a function to translate WKB to the ESRI personal geodatabase format and then have an update trigger take the submitted pgeo shapefile record from ArcMAP and translate it to the geometry type for updating the underlying table.

Re: geom null

Chris Hermansen — 2008-11-28T17:18:46

Then if you can't add to a null geometry, you will probably be forced to digitize your places into another table and join the two after the data entry phase. Which means you will need identifiers on the geometry sub table that can be linked 1-1 with identifiers on the attribute sub table. You can use a correlated subquery to find rows in the geometry sub table that do not connect with the attribute sub table. ju wrote:

Re: geom null

Paul Ramsey — 2008-11-28T17:05:50

It's very much a client side issue, not a PostGIS issue. Ask the QGIS/gvSIG/uDig folks how they support your use case. P. On Fri, Nov 28, 2008 at 8:12 AM, ju espaces-naturels.fr> wrote:

Re: geom null

ju — 2008-11-28T16:12:04

thanks but my table contains data for 1500 geographical place we are going to digitize some will be point, other line and other polygon so i can not make a false object and then modify it. Chris Hermansen a écrit :

Re: geom null

Chris Hermansen — 2008-11-28T15:32:43

Interesting question. What if, instead of null, you put in a simple piece of geometry that you can modify? (a point or a line or a 3-point linestring or a rectangle)? ju wrote: