How does one find out what happened to the administrative boundaries for these two US counties. They used to be there.
Relation: Black Hawk County (1792935) | OpenStreetMap and Relation: Monroe County (1804311) | OpenStreetMap both still seem to be there. Josm thinks they are valid relations.
I am not sure why Nominatim does not find them.
Edit: For completeness, Nominatim is what you search with when you use the search box at osm.org , or search via nominatim.osm.org .
Edit 2: I found the first of these by searching taginfo, and the second by browsing wikipedia to find out where it should be and querying OSM (using the “?” button) until I found it.
Yes, Rick (and hello and welcome), as Andy says, they are there. Is it “Nominatim only” where you do not see them? Why do you believe they “used to be there” when “they are there?”
If it is “Nominatim only,” you might want to directly contact @lonvia (Sarah Hoffmann, Nominatim’s author herself). Sarah, if this is “too public” a method to mention contacting you for this, please accept my apologies in advance; I’m simply trying to facilitate a question being answered. And again, Rick, please only contact Sarah if it is “Nominatim only.”
I’ve seen (using layers.openstreetmap.fr - Tuiles pour contrôle de données / Tile service with dataa checks ) “patchy holes” or “a county seems to have disappeared” before. This renderer doesn’t update quickly, but it looks like there is a “hole” in Arizona and another in Iowa (one county each). Sometimes I’ll take a look at OSM’s data and it is wrong and I’ll correct it, sometimes I can’t figure out what’s wrong (or it looks like nothing, or simply the renderer hasn’t caught up) and I’ll leave things alone.
well those two counties were not in the latest extract of OSM Boundaries. and when I log into OSM itself and search for “Black Hawk County Iowa” and “Monroe County New York” there are no results. Their boundaries were in the previous extract we got of OSM boundaries and now they are not. Hence, they used to be there. Can you find them in OSM ?
Both relations are missing the tag boundary=administrative, which probably explains why nomatim and OSM Boundaries don’t recognise them as boundary relations (anymore). I haven’t checked when or why this tag was removed in the history.
This is the CS which deleted the relevant tag in one of the counties.
Can someone help and add the tag back in so that they will appear again ?
For info about the changes see https://osm.mapki.com/history/relation/1792935 and https://osm.mapki.com/history/relation/1804311.
Edi: I’ve commented (in a hopefully friendly way) on the changesets.
I’ve done that - with a link to this topic in the changeset comment.
It appears that someone has fixed this. I see both in OSM now. Thank you!
Thanks for the feedback, Rick.
Glad to help… see my other ticket about Kenai Peninsula Borough and Lake and Peninsula Borough in Alaska. They are in OSM but didn’t show up in the extract of the US Admin Boundaries.
Sorry, Rick: I don’t know your context for “other ticket.” I’d be happy to look if you offer a link or similar.
I looked at tags on KPB and L&PB and they seem OK (as of now, even as you say they are in error). I compared them to neighboring Kodiak Island (Borough), which you don’t seem to indicate has problems (or are part of your “ticket”): tags are similar. KP has a
population=58059 tag and the others mentioned don’t, but that shouldn’t make a difference.
It isn’t always tags on a (mult)polygon that can make it wrong: other things can go wrong, usually something having to do with broken geometry. Incidentally, I looked at the Black Hawk County recent edits (up to its Version 11, edited by @Vinzenz_Mai less than 24 hours ago, as opposed to Version 10, which was in error) and indeed, @alphensebezorger is correct: the
boundary=administrative tag was missing, now fixed.
What “extraction” of US Admin Boundaries are you using / or how are you doing that? It is not obvious to me why these might be excluded, but perhaps with more info we can better sleuth as to why.
thanks. Well I am not the one doing the extrancting so I can’t say. It’s done by a group called Project-GC (where the GC stands for Geocaching) and they put the boundary data they obtain onto a server and make it available for others… it’s OSM-Boundaries.com which is where I dowload the info for the US. anyway, just the four US counties were missing. We now know why Monroe County NY and Black Hawk County Iowa were missing. that leaves the other two Alaska Boroughs. I will say Kenai Peninsula was fine a couple of years ago when they did the extraction, but now it’s not in the current file. Lake and Peninsula was missing a few years ago too so whatever is wrong with it - it’s still wrong. I found this tool, http://ra.osmsurround.org/index which thinks both relations 2605325 and 2605288 are ok. so I don’t know what the issue is. maybe it’s ok and given a different name in the the file???. Like, one I found. Washington County Rhode Island, appears as “South County” in the extract …
If we look at that: Relation: South County (2554903) | OpenStreetMap
official_name=Washington County and
The name was changed about two years ago.
Lake and Peninsula Borough https://www.openstreetmap.org/relation/2605288:
official_name=Lake and Peninsula Borough name=Lake and Peninsula
But this is valid since the start more than 10 years ago.
The same is valid for Kenai Peninsula:
official_name=Kenai Peninsula Borough name=Kenai Peninsula
Maybe you should look at
Lake and Peninsula and
Kenai Peninsula instead.
I did. nothing in the file with those names (either with or without the “Borough”). Very odd.
Sorry, Lake and Peninsula is missing in OSM-Boundaries.com right from the start.
Kenai Peninula was listed in the database November 2021 for the last time. The last two changesets were made in this month.
I can’t tell you why
The “why” might or might not “arrive” (as knowledge) in the future. Right now, I’d say “nice sleuthing” (again, too) to everybody. Nice to see effective chit-chat and enough trust and links and good will and things pointing in the right directions, etc. Gears do turn.
Rick, I’d say that your OSM-B dot com source “lags in time” a certain bit (weeks? more? neither would surprise me). And you (sure you do!) expect your source data to be more realtime than they actually are.
I don’t want to say “case closed” as there may be more that reveals as to “why,” I’ll simply say “door open.” Nodding my head here, I think I see others nodding our heads, too.
well the data I pulled says it’s from Jan 2, 2023. There is a new one in process that says from Feb 6, so we’ll see what it ends up having. I expect those two Alaska Boroughs to be missing still, but we shall see.
Thanks for the info and help!
As someone who has decades of experience of this sort of “not quite realtime lag” (sometimes hours or days or weeks or months), you are very welcome!
Adjust expectations, “get with the rhythm” (we even have a kind of rhythm with national bike routes and state DOTs and AASHTO and how our Cycle layers render, in the USA) … there is a certain amount of “let the renderers catch up.” Rinse, repeat. Good old OSM, yet again!
This was kind of neat. Applause to @Vinzenz_Mai especially, really, everybody here who participated or even simply “read along” with the history line!