As part of another project, we have deleted duplicate ways from a local copy of an OSM database. In some cases, these duplicates may be legitimate, but in other cases, are an error (e.g. way 481885892 and 55546011). I will try and fix some by hand, but is there a good way to tag or a place list the potential duplicates for others to inspect? Adam asked 02 Dec '17, 23:07 amb_santacruz |
How many are we talking about? If it's just a few, it would be better to just change them by hand. If it's a lot, you could make a map roulette challenge with the problem ways. OSM doesn't really have a single place you can report bad data to. Since it's a wiki, the idea is that you would fix it yourself, rather than reporting it to someone. If the bad data is just in one area/country, you could try contacting the local community in that area/country, they might know the source of the problem and might be interested in fixing it up themselves. answered 03 Dec '17, 13:07 rorym Thanks for the map roulette idea. Yes, it's a lot (> 100,000 potential duplicates, worldwide).
(03 Dec '17, 17:07)
amb_santacruz
1
whelp! >100k sounds very weird. You should probably post to the osm talk mailing list ( https://lists.openstreetmap.org/listinfo/talk ), or make an OSM diary entry. That's a lot of data.
(03 Dec '17, 20:28)
rorym
Its been mention that htis might have been due to an issue with the PT plugin, but without some more concrete information, it is difficult to determine if it was "only" that, or if there are other clear reasons.
(04 Dec '17, 17:09)
SimonPoole ♦
I will identify some more examples.
(05 Dec '17, 04:49)
amb_santacruz
|
I guess Osmose already covers it, more specifically the duplicate geometry error. An alternative is OSMInspector's duplicate ways Both error lists are updated automatically, so there is no need to report them, we just need someone to fix them :) answered 04 Dec '17, 12:59 escada Thanks for the links. I'm curious why these weren't picked up by OSM Inspector nor Osmose, as Stereo points out. I couldn't find detailed documentation, but perhaps these don't account for partial duplicates? For example, where the first segment of way A is duplicated by way B, but the geometries are not identical because way B is shorter.
(05 Dec '17, 04:49)
amb_santacruz
|
That particular way, I think, was due to a bug in the PT Assistant plugin that has now been fixed. It's interesting to note that neither osmose nor osmi spot it. answered 04 Dec '17, 17:08 Stereo |