Simon Cox writes:
I have some site lists that I want to use in GRASS.
Getting them in is not a problem, but then it looks
like you are on your own. Are there any buffering/masking
operators for site data, or do you have to convert to rasters
using s.menu?
As far as I know, there is no direct buffering technique for sites - you
can take the neighbourhood option in s.menu as a start, but I found that
not very satisfying. I usually work the other way around, and buffer the
map layer that you want to use for an analysis, and then go on with
s.menu. I did however make the program that Darrell mentioned to make
circles (well, circles... let's say circe-like lines), and that could be
modified to produce circular areas. But my experience is that buffers
around sites tend to overlap, so they won't produce data per site. So
what I actually do is the following: make a script that reads the
site_list line by line and takes each coordinate as the input-coordinate
for a r.los command that has as input-layer a map that has no 'relief'
with a maximum distance that would be equal to your desired buffersize.
This produces lots of small maps that can be used as a mask for r.stats
or r.report. I can send you an example if you like.
And, how to cope with multiple attributes for each site?
It looks like this needs multiple site lists? Am I right?
Last time I tried to get rim and s.db.rim working I wasted
a few days with no reward. What do other people do?
Yes, you'll have to use multiple site_lists if you're not very familiar with
awk. The grass-informix tools work as well, but then you need to buy
Informix. Anything, including rim, that helps you to select from a
site_list will be helpful. By the way, to my knowledge the concept of
'attribute of a site_list' is not very clear in GRASS. Some programs (like
s.surf.tps) assume that you have a 'real' attribute when its preceded by a
#, and otherwise it's considered a comment. For screen display: I found the
d.what.site program (I don't remember who wrote it, it's somewhere on the
net) very helpful, and I'm sure I'm not the only one that has made a
program that can make a label-file from a site_list so you can actually
SEE what's in you site_list.
Philip Verhagen
P.S. Concerning your 2nd questiom about the coordinates and r.mapcalc: I
liked Malcolm's suggestion very much. Like you, I already made plans to
use r.stats for a time-consuming shell-script...
--
S t i c h t i n g R A A P
Regionaal Archeologisch Archiverings Projekt
adress: Plantage Muidergracht 14
1018 TV Amsterdam
THE NETHERLANDS
phone/fax: (31) 20 525 5834
e-mail: motte@hacktic.nl