[Talk-us] mapping job - unconnected _link roads

[Martijn van Exel]

One thing I do not understand is why the script catches the false
positive node 560176247 [1] (just an example, there are more like this
one).
Anyone with some SQL savvy care to look into that?

[1] http://www.openstreetmap.org/browse/node/560176247

On Tue, May 29, 2012 at 9:27 PM, Martijn van Exel <m at rtijn.org> wrote:
> The example I gave isn't great - these _link roads are actually
> connected, they just dangle. But there are lots that are actually not
> connected. It looks OK on the map usually, and it's easy to miss when
> you're editing, but this type of error messes up routing.
>
> Martijn
>
> On Tue, May 29, 2012 at 9:24 PM, Martijn van Exel <m at rtijn.org> wrote:
>> Hi,
>>
>> I've been looking at unconnected _link roads (see here [1] for a
>> typical example) for the US, and found 1200 likely candidates. A
>> shapefile as well as an .osm file you can load in JOSM can be
>> downloaded here: http://lima.schaaltreinen.nl/osm/
>> The script I ran (on osmosis snapshot + linestring schema) is here:
>> https://github.com/mvexel/osm-gradesep/blob/unconnectedlinknodes/sql/09-missinglinknodes.sql
>> There are some false positives in there, but it looks like more than
>> 90% of these nodes warrant inspection.
>> I ran this analysis on the last planet from before 1/4, so some of
>> this may already have been fixed.
>>
>> [1] https://www.dropbox.com/s/85dqb7fm7whevxw/unconnectedlink.png
>> --
>> martijn van exel
>> http://oegeo.wordpress.com
>
>
>
> --
> martijn van exel
> http://oegeo.wordpress.com



-- 
martijn van exel
http://oegeo.wordpress.com