[R-sig-Geo] problem with extract() raster package using weight=TRUE

[up]

Robert Hijmans wrote
> Umberto,
> 
> The work-around is to assure that the CRS are the same:
> 
> projection(regioni) <- projection(t500)
> 
> You should of course only do that when the projections are in fact the
> same, as in this case. The problem is caused by sp::identicalCRS that
> compares the verbatim, but not the semantic representation of the CRS
> (and, therefore, perhaps should give a warning rather than an error?).
> 
> Robert

Many thanks Robert for your suggestion! :-)

But I've already tried that, and did not work.
Yestarday I tried to use another shapefile, because I suspected something
wrong in my WGS84 shapefile.
And with another shapefile things gone well!

I've downloaded a shapefile of my country but in a different projection
("+proj=utm +zone=32 +ellps=intl +units=m +no_defs"), and then reprojected
in "+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0" with:

liguria<-spTransform(ligu_utm, CRS("+proj=longlat +datum=WGS84 +no_defs
+ellps=WGS84 +towgs84=0,0,0"))

Then I've extracted values with the same function:
estra_liguria<-extract(prec,liguria,weights=T,cellnumbers=T,df=T)

and that's the result:

   ID cell totprec weight
1   1   10       0   0.12
2   1   18       0   0.34
3   1   19       0   0.33
4   1   20       0   0.42
5   1   25       1   0.57
6   1   26       0   0.60
7   1   27       0   0.84
8   1   28       0   0.96
9   1   29       0   0.91
10  1   30       0   1.00
11  1   34      10   0.06
12  1   35       5   0.84
13  1   36       0   1.00
14  1   37       0   1.00

I don't know why the first shapefile doesn't work (I've downloaded it from
http://www.gadm.org, lonlat WGS84), but the new one does the job! :-)

In any case, thanks very much!

Umberto




--
View this message in context: http://r-sig-geo.2731867.n2.nabble.com/problem-with-extract-raster-package-using-weight-TRUE-tp7584665p7584673.html
Sent from the R-sig-geo mailing list archive at Nabble.com.