[R-sig-Geo] write.nb.gal doesn't include correct ID

[Roger Bivand]

On Thu, 8 Jul 2010, Chris Fowler wrote:

> I am trying to write a rather large neighborhood list to a .gal file using 
> the following code
> write.nb.gal(Q1.nb, "Map 
> Data/subWeights.gal",oldstyle=FALSE,shpfile="submap.shp",ind="ind")
>
> where ind is an array of numbers representing FIPS codes at the census tract 
> level (I also tried it using "STFID" which is the same data as a column in 
> the shapefile submap.shp) while I don't get any errors, my .gal file uses the 
> sequential id number from the shapefile (a meaningless 1 to n number assigned 
> by ArcGIS in the creation of the shapefile). I have tried (in ArcGIS) 
> converting the number assigned by ArcGIS to my FIPS codes using Field 
> Calculator, but this doesn't seem to change anything. My concern is that the 
> problem is perhaps the length of the FIPS codes which are 10 or 11 digits 
> each?
>
> Is there something that I am missing that will allow me to use the more 
> meaningful FIPS codes in the gal file?


You must include more of your workflow. I suspect that the correct values 
are not being set when the "nb" object is created (synax and argument 
names may vary between functions). Please also explain which other 
software you are using - ArcGIS is by no means available to everybody, I 
for example have no access to this product.

Work from a canned example, such as:

library(spdep)
example(nc.sids)
nbs <- poly2nb(nc.sids, row.names=nc.sids$CNTY_ID)
writeSpatialShape(nc.sids, "nc_sids.shp")
write.nb.gal(nbs, "nc_sids.gal", oldstyle=FALSE, shpfile="nc_sids.shp",
   ind="CNTY_ID")

and explain again what the problem is. By the way, Arc typically assigns a 
FID which is 0-based, the 1-based indices are the default in functions 
making "nb" objects when no alternative is given, so your analysis is more 
likely guessing than actually what is happening. Note that you may also 
run into problems with apparently numeric identifiers (like some FIPS) 
that begin with "0", you'll have to feel your way there.

Hope this helps,

Roger

>
> Thanks,
>
> Chris Fowler
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no