[Tagging] admin_level on nodes: wiki vs practice

Martin Koppenhoefer dieterdreist at gmail.com
Tue May 13 12:50:00 UTC 2014

2014-05-13 14:20 GMT+02:00 Ilya Zverev <zverik at textual.ru>:

> admin_level on place nodes surely duplicates admin_level tag value
> from one of relations which contain that node, but is that a bad
> thing?
> Did you try to calculate admin_level for a place in osm2pgsql
> database? I've spent two hours now trying to construct and optimize an
> SQL query for that, and seeing it takes at least 20 seconds for a
> tile, I'd prefer having admin_level tag on places.

admin_level has no real "definition" in the wiki what it is supposed to
express: the key link redirects to boundary=administrative:

This could already be taken as a statement: admin_level is there to express
the admin_level of an administrative entity (this text is not there in the
wiki right now).

IMHO "place"s are orthogonal to administrative entities. They can sometimes
cover the same area, in other cases they don't. Some places are
administrative centers for an administrative entity, others aren't.

Now there is also a key "capital" that can tell the administrative
importance for a place (it will contain the admin_level of the highest
administrative entity (=lowest admin level number) for which a place is the
administration centre), so no real need to have an admin_level with
duplicating values on these places as well. I think capital is a better key
for places then "admin_level", as any place will "have" lots of
admin_levels (e.g. 2 when it is inside a country) so semantically it
doesn't make a lot of sense.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tagging/attachments/20140513/cc6cf1ab/attachment.html>

More information about the Tagging mailing list