-----Ursprüngliche Nachricht-----
Von: Maciej Sieczka [mailto:tutey@o2.pl]
Gesendet am: Mittwoch, 16. August 2006 17:58
An: Praktikant/in2 ARP
Cc: 'grassuser@grass.itc.it'
Betreff: Re: AW: [GRASS-user] v.in.ascii error: G_reallocPraktikant/in2 ARP napisa?(a):
Thanks a lot for your help Maciek.
I`ve done everything as described. But when I run
nnbathy -W 0 -P alg=nn -n 300x300 -x 605450 606050 -y 232550 231950 -i
dtm >
interpolated_xyz_output
I get an error: Speicherzugriffsfehler (I think its something like
segmentation fault in English)Maybe there is an error in my input line?
I also tried r.in.xyz and r.surf.nnbathy.
r.in.xyz is working fine. But when I use the created raster I get the
error:
r.stats: 100%
nnbathy IN ACTION - MAY TAKE SOME TIME
PLEASE STAND BY UNTIL 'ALL DONE' IS PRINTED
CONVERTING nnbathy OUTPUT TO GRASS RASTER
ERROR: Data conversion failed at row 1, col 1
Sorry that I havn't answered for some days. Had some other things to do.
I'm thinking too that nnbathy is failing with the big text file. If I use a
small text file it's working.
But there's still the problem with r.surf.nnbathy which I would like to
use.It's always the same error whatever size the input file has.
That suggests nnbathy failed, then r.in.ascii couldn't read it's output.
(I should implement some error trapping here, hopefully will do; noted
down.)Where there any errors when you built the nn library or nnbathy?
No there were no errors
How big is your your input dtm text file? Can you send it to me (if not
more than few MB packed) or put somewhere for download?
It has about 300MB.Think thats a little bit too big for sending per mail.
Another problem is the licence so you can't download it.But I send you the
small text file which does work with nnbathy but doesn't work with
r.surf.nnbathy after importing in GRASS with r.in.xyz
Maybe something is wrong with the text file. Or I make something wrong when
importing the xyz file in Grass.
Hope the text file helps.
Thanks a lot
Best regards
Michael
I just tried reproducing your problem, as far it is possible with
limited information you provided, and nnbathy 1.66 as well as
r.surf.nnbathy work OK for me. I need your data to test further.Maciek