<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'><div dir='ltr'>
On error, a problematic point is indicated, but it refers to the index<div>being developed by splitting and doesn't refer to the original, offending row oid</div><div>of the source file being processed to help tracing. </div><div><br></div><div>Sometimes a geom or geom range is returned that I tried to use </div><div>previously to focus in on the possible problem, unsuccessfully thus far.</div><div><br></div><div>Attached a file which may or may not be of interest, two smallish</div><div>shape files that successfully became topologies independently,</div><div>but failed when I attempted to combine them. One multi-ploygon,</div><div>one linestring. </div><div><br></div><div>A log of my approach is included, which is probably the source of</div><div>error.</div><div><br></div><div>I will address myself to noding and St_<functions, with tolerances></div><div>to see if I can clean better before topology.</div><div><br></div><div>Chris<br><br><div>> Date: Thu, 8 Dec 2011 09:17:53 +0100<br>> From: strk@keybit.net<br>> To: postgis-users@postgis.refractions.net<br>> Subject: Re: [postgis-users] Preparing for Topology - St_CreateTopoGeo<br>> <br>> On Tue, Dec 06, 2011 at 08:33:53PM -0500, Chris English wrote:<br>> > <br>> > In your experience, it there a method to predict where failuremight<br>> > occur, or some characteristics of offending geometries...<br>> <br>> Lines which are almost coincident but not quite.<br>> <br>> > I was just wondering if through the multitudinous process of st_split,<br>> > addface, addedgemodface,if a progression of small remainders from the<br>> > splitting (or others) get carried forwardthat would fault st_createtopogeo<br>> > as it gets around to updating {left,right}_face fields and the like.<br>> <br>> The function performs noding as first thing so maybe you could do the<br>> noding yourself on the whole data and see if it still creates problems.<br>> If it does you could then select only the (noded) edges that intersect<br>> a small buffer around the reported problematic point (was such problematic<br>> point reported ?).<br>> <br>> --strk;<br>> <br>> ,------o-. <br>> | __/ | Thank you for PostGIS-2.0 Topology !<br>> | / 2.0 | http://www.pledgebank.com/postgistopology<br>> `-o------'<br>> <br>> _______________________________________________<br>> postgis-users mailing list<br>> postgis-users@postgis.refractions.net<br>> http://postgis.refractions.net/mailman/listinfo/postgis-users<br></div></div> </div></body>
</html>