I have a database of 50,000 or so villages and towns in Burma. Quite a few have two or more names, and the database copes with this by putting two or more points on the same spot, each with a different name. I am trying to do nearest neighbour analysis on clusters of sites, using two copies of a file of selected data. But if I select villages within a certain area, I get a lot of duplicates, resulting in a statistically unviable neighbour to neighbour value of 0 for each duplicated site. I guess one possibility is to select my sites within the required area, then manually create a single new point on each one in a new layer, just for the purposes of NN measurement, but is there a less time-consuming way to do it?
The database, by the way, is from the US Geonet names server, a wonderful free resource, and God bless those American taxpayers who have provided it :-) http://earth-info.nga.mil/gns/html/cntry_files.html Bob Hudson Archaeology Department, University of Sydney, Australia. http://acl.arts.usyd.edu.au/~hudson/bobhpage.htm Using MI 7, Win XP Pro --------------------------------------------------------------------- List hosting provided by Directions Magazine | www.directionsmag.com | To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] Message number: 11138
