[GRASS5] r.profile

[Hamish]

> > after toying around with the enhanced version of r.profile (great job,
> > Bob!) I wonder if there's an _elegant_ way to pipe any output from
> > v.out.ascii into r.profile.  
> >   
> > This is trivial for vectors consisting of just one line, but for
> > multi-line vectors manual editing is still needed.
> 
> I wouldn't call it elegant, but:
> 
> # lines to e1,n1,e2,n2[,...]
> v.out.ascii multi_line format=standard | sed -e 1,10d | \
>    grep -v ' 1 ' | tr '\n' ',' | sed -e 's/,L/\nL/g' | \
>    cut -f2- -d',' | sed -e 's/^ //' | tr ' ' ',' | \
>    sed -e 's/,$//' > ascii_lines.tmp
> 
> for LINE in `cat ascii_lines.tmp` ; do
>    echo "$LINE" | r.profile -gc in=raster_map
> done
> 
> 
> I am sure awk would be prettier.


Now that CVS is back I can test...


Correction:  piping commas from stdin causes a segfault


v.out.ascii multi_line format=standard | sed -e 1,10d | \
    grep -v ' 1 ' | tr '\n' ',' | sed -e 's/,L/\nL/g' | \
    cut -f2- -d',' | sed -e 's/^ //' | tr ' ' ',' | \
    sed -e 's/,,/,/g' | sed -e 's/,$//' > ascii_lines.tmp


g.region rast=raster_map
LINE_NUM=1
for LINE in `cat ascii_lines.tmp` ; do
   r.profile -gc in=raster_map profile=$LINE > line_${LINE_NUM}.prn
   LINE_NUM=$(($LINE_NUM + 1))
done



CVS updated to allow comma separated pairs from stdin, e.g.:
e1,n1
e2,n2
[...]
eN,nN


but 

echo "e1,n1,e2,n2[,...,eN,nN] | r.profile

will still segfault.



Hamish