[Imports] NHD Duplicate data

Mike N. niceman at att.net
Thu Sep 10 16:29:41 BST 2009


Thanks - that sounds like a good technique that I could use in the future to catch pre-mapped areas.

   A bit of background - the National Hydrography Dataset is divided into river basin and sub-basins.   The sub-basins are a manageable size to edit if necessary and upload.   Right now, active mappers are interested in uploading the NHD data in their areas because it's instinctive for new mappers to map lakes and rivers.   That's a waste of their valuable time since the NHD data is better than Yahoo aerial tracing in most cases.  So there are multiple people performing data imports for different areas.

   Back to this problem, the default NHD scripts add the NHD database ID for each water body (ComID), and this simplifies finding duplicate water bodies - I can compare ComID's instead of collections of nodes, ways, and polygons.   That search can even be done for data that is already uploaded, except it is greatly complicated if someone has edited one or both of the duplicates.   

  I'd rather not soak up system time and space importing something that's a duplicate.   So I'll add a note to the NHD Wiki page about this and crosscheck all the data I've converted.




From: Emilie Laffray 
Sent: Thursday, September 10, 2009 8:57 AM
To: Mike N. ; imports at openstreetmap.org 
Subject: Re: [Imports] NHD Duplicate data


    I'm not sure what the best solution is; a program to detect duplicate
  areas may not detect duplicates from an adjacent upload by a different
  person.



For the Corine import of France, I have been using Postgis and some SQL to calculate overlap between two surfaces. It is a mixture of ST_Area with ST_Intersect. It is not perfect but we caught lots of existing forests inside OSM so we didn't have to upload them. 


I am not exactly sure what your problem is but we could have a look.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/imports/attachments/20090910/4f162b8d/attachment.html>


More information about the Imports mailing list