[GRASS5] [bug #858] (grass) v.support/v.build bug identified!

[Radim Blazek]

On Mon 26. November 2001 20:03, Request Tracker wrote:

I'm not sure if this is realy v.build bug. v.build is designed for building
topology from clean linework. Two polygons with duplicated boundary is
not clean. Unfortunately cleaning module v.spag is only reliable for '-i'
flag. 
My solution for those who have imported files with clean topology
(export from ArcInfo for example, not ArcView) and good computer 
(time consuming method):

v.drop input=bad output=bad.drop   
cp vector/dig_att/bad vector/dig_att/bad.drop  # sorry, no support for dig_att files in v.drop
v.support map=bad.drop
v.spag -i map=bad.drop
v.build.polylines -c -o input=bad.drop output=bad.ok type=area

Result:
 Number of lines:   3
 Number of nodes:   2
 Number of areas:   2
 Number of isles:   1
 Number of atts :   2
 Number of unattached atts :   0
 Snapped lines  :   0

This should reduce size of the vector (areas) to one half also.

v.drop: http://www.raz-dva.cz/krakonos/v.drop.tar

Radim


> this bug's URL: http://intevation.de/rt/webrt?serial_num=858
> -------------------------------------------------------------------------
>
> Subject: v.support/v.build bug identified!
>
> grass downloaded at: CVS Server, Germany
>
> The v.support/v.build bug with causes the PNT_TO_AREA failed warnings
> seems to be identified.
>
> A micro-demo-location to reproduce the problem is available at
> http://mpa.itc.it/markus/tmp/
>
> The location contains to sample vector maps "good" and "bad",
> each with two polygons only:
>
> GRASS:~/grassdata/patdemo > l
> total 12
> drwxr-xr-x    2 neteler  ssi          4096 Nov 26 19:49 PERMANENT
> drwxr-xr-x    2 neteler  ssi          4096 Nov 26 19:52 rawdata
> drwxrwxrwx    7 neteler  ssi          4096 Nov 26 19:55 vector
>
> In rawdata are the ARC-ASCII vector files and two import
> scripts:
>
> GRASS:~/grassdata/patdemo > ls -la rawdata
> total 40
> drwxr-xr-x    2 neteler  ssi          4096 Nov 26 19:52 .
> drwxr-xr-x    5 neteler  ssi          4096 Nov 26 19:51 ..
> -rw-r--r--    1 neteler  ssi            35 Nov 26 19:51 BadExample.dat
> -rw-r--r--    1 neteler  ssi            67 Nov 26 19:51 BadExample.pnt
> -rw-r--r--    1 neteler  ssi           273 Nov 26 19:51 BadExample.pol
> -rw-r--r--    1 neteler  ssi            35 Nov 26 19:51 GoodExample.dat
> -rw-r--r--    1 neteler  ssi            67 Nov 26 19:51 GoodExample.pnt
> -rw-r--r--    1 neteler  ssi           273 Nov 26 19:51 GoodExample.pol
> -rwxr-xr-x    1 neteler  ssi           154 Nov 26 19:52
> import.BadExample.sh -rwxr-xr-x    1 neteler  ssi           159 Nov 26
> 19:52 import.GoodExample.sh
>
> Now the GRASS bug: v.support writes
> PNT_TO_AREA failed: (1681015.000000, 5118635.000000) (Category 694020)
> PNT_TO_AREA failed: (1680445.000000, 5118795.000000) (Category 54014)
>
> when two or more polygons have the same starting point (the "bad" map).
> The identical "good" map with rotated nodes runs well.
>
> That may be the key to understand all the E00 and SHAPE problems
> which are regularly reported.
>
> Thanks to Lorenzo Potrich (ITC-irst) for identifying this problem.
>
> Please someone may fix this severe problem...
>
> Markus Neteler
>
>
> -------------------------------------------- Managed by Request Tracker
> _______________________________________________
> grass5 mailing list
> grass5 at grass.itc.it
> http://grass.itc.it/mailman/listinfo/grass5