Hi Markus,<br><br>I ditched osmosis because I wanted to know where they were being dropped so wrote my own filter tool which gave the same results. Looks like it's just how it is, but I'm wondering what the point of a way with broken noderefs is. Is it corruption or a feature?<br>
<br>David<br><br><div class="gmail_quote">On Tue, Oct 16, 2012 at 5:33 PM, <span dir="ltr"><<a href="mailto:marqqs@gmx.eu" target="_blank">marqqs@gmx.eu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi David,<br>
<br>
did you find out exactly where in your toolchain these nodes get lost?<br>
<br>
Maybe it's the filter process with Osmosis? I wouldn't assume a bug, I think it's more likely a misunderstanding in how the program options should be used.<br>
<br>
I'm not sure about the official API definition but it might be perfectly normal that not every noderef has a corresponding node in the file.<br>
<br>
If you do cheap extracts (e.g. with osmconvert without "--complete-ways" option) the way objects will not be modified. Each way object will stay complete, even if all of its nodes (but one) has been removed from the file.<br>
<br>
Regards<br>
Markus<br>
<br>
-------- Original-Nachricht --------<br>
> Datum: Tue, 16 Oct 2012 16:40:18 +0100<br>
> Von: David Prime <<a href="mailto:david@primefarm.co.uk">david@primefarm.co.uk</a>><br>
> An: <a href="mailto:dev@openstreetmap.org">dev@openstreetmap.org</a><br>
> Betreff: [OSM-dev] missing nodes<br>
<div class="HOEnZb"><div class="h5"><br>
> Hi All,<br>
><br>
> I'm relatively new to processing the OSM datasets and trying to extract<br>
> interesting data. What I'm trying to do is build a dataset of all UK<br>
> roads.<br>
> I first used osmosis with a way filter based on the highway tag, which<br>
> worked reasonably well. However, when I delved into the data, I noticed<br>
> that a lot of my ways had node refs to missing node entities in the file<br>
> I'm extracting data from (a recent england dump). Now, I expect some roads<br>
> to get chopped off at borders and so on but I've done it manually via a<br>
> java program that parses the full england dump and I've counted<br>
> ~1.3million<br>
> ways which reference nodes that I could not find in the same XML file.<br>
> This<br>
> seems excessive to me, unless I'm misunderstanding some aspect of the data<br>
> model. Any ideas from the old hands?<br>
><br>
> Many thanks,<br>
> David<br>
</div></div></blockquote></div><br>