Dear All,
I am using Wingrass generic in a cygwin environment/windows xp. After a simple processing of raster maps with r.mathcalc, the resulting maps are about 17times larger (from about 28 to about 510 MB in the fcell subdirectory of the mapset) than the original maps (the original data are 2800, 2900 etc.; the resulting data are 25.5, 29.8 etc). Does a possibility to compress the resulting data? Thanks, Jan Bartholdy
Jan Bartholdy wrote:
I am using Wingrass generic in a cygwin environment/windows xp. After a
simple processing of raster maps with r.mathcalc, the resulting maps are
about 17times larger (from about 28 to about 510 MB in the fcell
subdirectory of the mapset) than the original maps (the original data
are 2800, 2900 etc.; the resulting data are 25.5, 29.8 etc). Does a
possibility to compress the resulting data? Thanks, Jan Bartholdy
There are several possible reasons for expansion, including:
a) the result was generated at a higher resolution than the input maps.
b) the input maps are compressed but the output maps aren't.
c) the input maps are 1- or 2-byte integer maps but the output maps
are 4- or 8-byte floating-point.
For a), set a lower resolution with e.g. "g.region res=...".
For b), use r.compress.
For c), explicitly coercing the data with int(...) will use 4-byte
integers; AFAICT, there isn't any way to convert a map to 1- or 2-byte
integers.
--
Glynn Clements <glynn.clements@virgin.net>