Glynn Clements wrote:
Rainer M. Krug wrote:
I have a raster which has very large cell values (>10^10) which I can
not store in the grid directly.
Do you have to use integers? A DCELL (double-precision floating point)
map can store integers up to 2^53 (~10^16) exactly.
No - I don'y have to use integers - DCELL would be fine as well. but how
can I change the format? The raster is created from an imported shape file.
In a second step I want to use r.reclass() to create the second raster
which holds the large cell values. Can I specify in r.reclass() that the
cell values should be stored as DCELL?
Therefore I have at the moment category
values of 1-6 and 1 represents 10^10. As I want to use mapcalc with
these values, I want to use labels to store the large numbers in, but as
I have to do it for a large number of maps (which are generated during
simulations) I can't do it manually by using r.support interactively.
Therefore my question: is there a way to either store the values in the
grid cell itself instead of the category value or if not, how can I
create these labels non-interactively?
The cats file format is simple enough that you could generate it from
a script. See the cats files for the sample datasets (e.g. Spearfish)
for examples.
Generating a cats file using the libgis functions is also quite
straightforward e.g.:
const char *map_name = "outmap";
const char title[] = "title string";
const char *labels[NUM_CATS] = { ... };
struct Categories cats;
G_init_cats(0, title, &cats);
for (i = 0; i < NUM_CATS; i++)
G_set_cat(i, (char *) labels[i], &cats);
G_write_cats ((char *) map_name, &cats);
Thanks - but as you suggested above, the DCELL solution appeals more to me.
_______________________________________________
grassuser mailing list
grassuser@grass.itc.it
http://grass.itc.it/mailman/listinfo/grassuser