[GRASS5] [aaime@libero.it: [GRASS-bugs] vector files: topologyrulesbroken]

[David D Gray]

Andrea Aime wrote:
> 
> Hi everybody,
> in fact I read only after submitting bug that v.spag detects
> line crossings, but... if I use v.spag over a vector file
> containing only two lines crossing I end up with a new vector
> file with lots of lines and nodes (much more than 4 lines and
> 5 nodes that I expected). I was still using grass5beta8, if
> that bug has already been corrected, I apologize... I'm
> downloading beta11, tomorrow I will made another try.
> Bye
> Andrea

Hi Andrea

I have noted before that v.spag likes the lines to cross in a
well-behaved manner. If lines run over one another, they are not split
or are not processed correctly, or the module bombs out. ie, if you
have:

X------------X------------------------X
X-------------------------------------X

where the lines are supposed to be co-incident, you get:

X--X-X---X-X

and so on, or other errors.

I suspect that these problems come mostly from importing
double-digitised data, but it is certainly a problem if you are getting
duplicate lines from manual digitising. 

I've sometimes found it helps a bit to run v.spag -i first (to do just
an identical lines check) then run the full check (without the -i
option). But v.spag -i cannot clean the example above, because tho'
these lines are co-incident, they are not identical. (Also this would
have to be re-run on the original)

We are planning for the future of vector processing that vector data
will be loaded up to a network structure before any kind of processing
is done. That will enforce a guarentee that all lines will be single,
and not run over each other (v.in.shape already uses such a structure).

Regards

David

PS. I've just checked v.spag on a `well-behaved' spaghetti map. It
splits OK, but doesn't label the newly created segments automatically. I
think we need some idea of a composite line to be built in dig_plus, a
bit like areas are composed from area edges and a point locator, but for
linear features..?

> 
> Markus Neteler wrote:
> >
> > Hi vector gurus,
> >
> > find attached a vector bug which seems to need our attention.
> > Maybe it's an v.spag issue?
> >
> > Thanks,
> >
> >  Markus
> >
> >   ------------------------------------------------------------------------
> >
> > Subject: [GRASS-bugs] vector files: topology rules broken
> > Date: Thu, 15 Feb 2001 15:36:43 GMT
> > From: aaime at libero.it ()
> > Reply-To: grass-bugs at geog.uni-hannover.de
> > To: grass-bugs at geog.uni-hannover.de, grass-bugs at intevation.de
> >
> > Below is the result of your feedback form.  It was submitted by
> >  (aaime at libero.it) on Thursday, February 15, 2001 at 15:36:43
> > ---------------------------------------------------------------------------
> >
> > subject: vector files: topology rules broken
> >
> > area: grass5beta11
> >
> > platform: Linux/Intel
> >
> > vendor: RedHat
> >
> > cputype: Intel (i486, i586, pentium ...)
> >
> > xwindows: Xfree 4.0.1
> >
> > windowsmanager: KDE 2.x
> >
> > tcltkver: -- select --
> >
> > grassserver: Hannover site
> >
> > grassversion: Linux/Intel binaries
> >
> > cvsyesno: -- select --
> >
> > compiler: -- select --
> >
> > bugreport: Documentation (GRASS programmer manual, "Vector Maps" claims that vector files are made of nonintersecting curves, and a "Topology
> > rule" says that arcs should not cross each other. That's a standard
> > planar enforcement rule but... it's not enforced! I have digited
> > a file with 2 crossing lines, run v.support and it's still saying
> > "2 lines, 4 nodes". v.digit displays only four nodes. So, GRASS
> > does not split lines at crossing, and it doesn't create a node in
> > the crossing, nor it refuses to accept that bogus file.
> >
> > name: Andrea Aime
> >
> > ---------------------------------------------------------------------------
> >
> > REMOTE_HOST: monet7-176.comune.modena.it
> >
> > ----------------------------------------
> > If you want to unsubscribe from grass-bugs
> > mailing list write to:
> > minordomo at geog.uni-hannover.de with
> > subject 'unsubscribe grass-bugs'
> 
> ----------------------------------------
> If you want to unsubscribe from GRASS Development Team mailing list write to:
> minordomo at geog.uni-hannover.de with
> subject 'unsubscribe grass5'

---------------------------------------- 
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo at geog.uni-hannover.de with
subject 'unsubscribe grass5'