[STATSGRASS] unfilled matrices with R and gstat in grass
Roger.Bivand at nhh.no
Thu Feb 15 12:27:55 EST 2007
On Wed, 14 Feb 2007, Roger Miller wrote:
> I realized just recently that my reply on this thread went to Edzer
> rather than back to the list.
> Thanks for your replies. The scripts are attached. Both the wrong and
> the right (but odd) results resulted from running the same script with
> different data sets and different Grass region settings.
> The scripts are run in sequence "startup" followed by "read" and finally
OK. Replace "startup" by calling gmeta2grd(), see the example on the
gmeta6 help page. That gets the SpatialGridDataFrame object, if you need a
SpatialPixelsDataFrame object, coerce using
I have no idea what you are doing in "read", all you need to do is call
readVECT6(), the rest is I think superfluous (with the possible exception
of overwriting the WKT-style CRS object with the GRASS-style). The CRS
issue can be handled as:
proj4string(rb) <- CRS(G$proj4)
I would not have written "map" anyway. If you show your users how to use
the command line directly, they'll be able to fish themselves from then
on, and variogram fitting does need interaction.
Please try to get this working _using the command line only_ until it
works, then simplify it until it cannot be reasonably simplified further.
Then you will have what you need to build a TUI/GUI (a tcltk one for Rcmdr
came past earlier this week ...) if it is worth the trouble (IMHO it
> I also attached the script "tile" that I used to construct the 16 maps
> that I could patch together to get the correct result.
> The workspace is rather large, but it can be downloaded from
> If that doesn't work then let me know and I'll try something else.
> Roger Miller
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
More information about the grass-stats