[Imports] How to merge data to keep object histories when doing small-scale import?

Andrew Guertin andrew.guertin at uvm.edu
Tue May 24 21:09:48 UTC 2011

I'm doing a small-scale (~250 buildings) import of the University of
Vermont. (I posted about this a while ago, it took forever to get the data.)

What is the best way to merge the data I'll be importing with the data
that's already there? For example, I have building outlines that are
much better than what's already there, but some buildings have tags
(amenity=library, tourism=museum) that my imported data doesn't.

Further, when changes are made to the university's data source, I'd like
to be able to re-import and not lose any changes people have made to OSM.

Finally, there's the matter of history, which I'd like to be able to
keep if possible.

I'm most familiar with JOSM for editing OSM data, but I've also used
Merkaartor and Potlatch. In none of them could I find a way to, for
example, replace all the nodes in one way with a copy of the nodes in a
different way, but keep the first way's id, history, and tags.

For reference, here's how I am working/would like to work with the data:
I'm given an AutoCAD file with building outlines. I run it through a
hand-written program heavily based on ogr2osm.py, which calls to some of
our internal databases and spits out a .osm file. I then take that .osm
file and open it in a GUI OSM editor and hand-merge our data with OSM data.


More information about the Imports mailing list