I have done some optimization in cleaning functions and build
process. Build and cleaning should be now faster and more reliable.
However, maybe that I introduced new bugs, especially in point
in polygon/area function and assigning of isles and centroids to areas
could fail.
v.in.ogr without '-c' option should now in theory create clean file
which builds without "Number of incorrect boundaries" (i.e. 0).
I am interested in data where it fails, v.build error= can
extract errors from the vector.
Radim
Radim Blazek said:
I have done some optimization in cleaning functions and build
process. Build and cleaning should be now faster and more reliable.
However, maybe that I introduced new bugs, especially in point
in polygon/area function and assigning of isles and centroids to areas
could fail.
v.in.ogr without '-c' option should now in theory create clean file
which builds without "Number of incorrect boundaries" (i.e. 0).
I am interested in data where it fails, v.build error= can
extract errors from the vector.
Radim,
Thanks a lot for these improvements. I can now import the file I had
problems with (the one we had a private exchange about), without hardly
any errors (only a few areas without centroid). This makes v.in.ogr much
more usable for me.
Moritz
On Wednesday 13 August 2003 13:43, Moritz Lennert wrote:
Thanks a lot for these improvements. I can now import the file I had
problems with (the one we had a private exchange about), without hardly
any errors (only a few areas without centroid). This makes v.in.ogr much
more usable for me.
Missing centroids is the bug in OGR, the way around is v.format+v.clean.
Please comment the bug here
http://bugzilla.remotesensing.org/show_bug.cgi?id=213
if you want it to be fixed.
I added warning for shapefile driver to v.in.ogr, but it seems that
OGR_Dr_GetName() returns NULL, so it is not printed.
Radim