[OSM-talk] Crazy ways after editing - A real argument for changeset data

Andy Robinson (blackadder) blackadderajr at googlemail.com
Wed Feb 6 12:46:31 GMT 2008


David,

I took a look at this area a little more closely with JOSM and I can see its
not straight forwards to correct because it's not just one way that got
moved but blocks of connected ways from what I can tell. Unfortunately they
now lie over the top of other exiting ways so its difficult to select the
ones effected so that they can be simply slid back. I think the only way to
sort it will be to evaluate the history data for the nodes in the whole
area. From this you can see which nodes (and hence ways) were moved at
around the same time instance. 

For instance if I look at two connected ways that are displaced (6331826 and
6397132) and look at the node history for some of their nodes  I can see
that their position was last moved at a timestamp around
"2008-02-04T04:32:24+00:00" they vary by some minutes in some cases so
perhaps it took a while for all the changes to update to the database.
Anyway, if a bit more checking was done it should be possible to automate a
rollback for all the affected data. Would need someone perhaps to help with
scripting that so perhaps David Hansen can assist with ideas.

Another option might be to use last week's planet dump and do a compare and
revert between the state last week and the state now. This will though mean
you loose data from the time of planet last week to Feb 04 if a delete and
re-insert is done.

Lets hope the folks that are looking at change set info via the API can find
us a long term solution to this problem as I'm sure its going to happen
again soon.

Hope this helps  move things along for you.

Cheers

Andy

>-----Original Message-----
>From: talk-bounces at openstreetmap.org [mailto:talk-
>bounces at openstreetmap.org] On Behalf Of SteveC
>Sent: 06 February 2008 11:11 AM
>To: David Muir Sharnoff
>Cc: OSM-Talk Openstreetmap
>Subject: Re: [OSM-talk] Crazy ways after editing
>
>David,
>
>There is also a talk-us list which may be of help in future.
>
>On 5 Feb 2008, at 18:46, David Muir Sharnoff wrote:
>
>> The area of Oakland, California I've been editing
>> (with Potlatch) suddenly changed in a very bad way.
>>
>> There are a whole bunch of ways (at least 50,
>> perhaps a lot more) that have had one or more
>> of their points move several miles south-east.
>>
>> A major street, Park Ave, seems to have
>> disappeared entirely.   Or, at least, I can't find
>> it any more.
>>
>> Edit at this location to see:
>> http://www.openstreetmap.org/?lat=37.81886&lon=-
>122.21229&zoom=16&layers=B0FT
>>
>> Some example ways that are messed up:
>> 6403309, Saint James Dr.
>> 6380125, Somerset Rd
>> 6365079, Pershing Dr
>> 6391271, Indian Rd
>> 6381464, Hampton Rd
>> 6374990, Excelsior Ave
>> 6359191, E 33rd St
>> 6390297, Everett Ave
>>
>> Does anyone have any idea what happened
>> or how to fix it?
>>
>> Thanks,
>> -Dave
>>
>> _______________________________________________
>> talk mailing list
>> talk at openstreetmap.org
>> http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk
>>
>
>have fun,
>
>SteveC | steve at asklater.com | http://www.asklater.com/steve/
>
>
>
>_______________________________________________
>talk mailing list
>talk at openstreetmap.org
>http://lists.openstreetmap.org/cgi-bin/mailman/listinfo/talk





More information about the talk mailing list