[Talk-us] Duplicate TIGER imports

[Chris Lawrence]

On Tue, Apr 14, 2009 at 5:41 PM, Minh Nguyen <mxn at 1ec5.org> wrote:
> A few weeks after DaveHansen imported the TIGER data for Ohio [1],
> RSatterf reimported the data for Greene County, Ohio [2]. As a result,
> every single way in the county is duplicated. On top of that, RSatterf's
> import also added a "loose" node (not connected to any way) for every
> node of every way in the county. Here's an example at the
> Montgomery–Greene county line [3]. It's a real mess, and it looks like
> we just need to delete every node or way last edited by RSatterf. We
> should keep ways that've been edited by another user since then, because
> in some cases DaveHansen's motorways etc. have probably been deleted in
> favor of RSatterf's. So does anyone have a handy mass deletion script?

It probably wouldn't be too hard to whip up something that uses the
extended API to query the database and generate an OSM XML file you
could churn through using the bulk uploader script (probably just an
XSLT transform of the returned XML would do it).  The only issue is
that you might have some shared nodes that are in-use by ways that you
don't want to delete.

After API 0.6 is up it would probably be a lot less of a pain to do, too.


Chris