[OSM-talk] Potlach problem?
david at frankieandshadow.com
Thu Jul 12 10:59:02 BST 2007
On 12/07/2007 10:05, Richard Fairhurst wrote:
> David Earl wrote:
>> However, this begs the question how ways and segments are still getting
>> into the database without the nodes the segments refer to.
>> I grepped the planet file and found it. It reports created_by='Potlatch
>> alpha' and has quite a few segments, from which I chose "27811754" at
>> random; grepping for that I get
>> <segment id="27811754" from="31657145" to="31657146"
>> (no created_by, and since JOSM generally puts one in, I assume it wasn't
>> created by JOSM), and from that I grepped for 31657145, the first node
>> reference, and found that this is not in the planet file.
> Hm. http://www.openstreetmap.org/api/0.4/node/31657145 looks ok to me:
Maybe it is the planet export then?
Do you think it could be that the export takes ages, so the offending
node(s) weren't in the database when the nodes were exported, but by the
time the dump reached the segments, someone had uploaded the new set of
nodes, segments and ways, so the segments and ways appear in the dump
but not the nodes?
If this is the case, there is the potential for all sorts of
inconsistencies and integrity problems in the planet file, not just
> <?xml version="1.0" encoding="UTF-8"?>
> <osm version="0.4" generator="OpenStreetMap server">
> <node id="31657145" lat="-37.9600081016474" lon="145.054282715186"
> visible="true" timestamp="2007-07-11T06:46:45+01:00"/>
> What's the way ID?
Since you gave me a lat/lon for the node, I just downloaded an area
centred there in JOSM. (It appears to be near Melbourne Australia, not
Regina Canada as I had guessed). It does all look OK.
I didn't check all the segments, I just chose one at random.
But I've just double checked and I can't still can't find any occurrence
of 31657145 in the planet file other than in the 'from' part of the
(BTW Sorry, the name was 'Southland Shopping Centre' (singular) not
'Southlands Shopping Centre' as in my original message)
I also just chose that Way at random to check. There are scores of
others that I think are the same problem.
>> Of course this combination of segments and nodes may have been modified
>> later, but the smoking gun in the file is 'Potlach alpha'.
>> Sorry if you already know about this.
> No problem. BTW, Potlatch has two 't's:
> http://www.google.com/search?q=potlatch+lettrist :)
I thought I'd copied and pasted that from the grep, but apparently I
More information about the talk