Polygons disappear after v.generalize simplification

[dido]

Polygons disappear after v.generalize simplification.

A large rectangular polygon dataset is used with ~14 000 features. Polygons are first processed with rmarea to remove all polygons of size 900 m2. Then a simplification is executed with v.generalize using a factor of 10, all other parameters defaulted.

After simplification several large polygons are missing from the set rendering it useless.

System is a win7x64; QGIS 2.6.1 and GRASS 6.4.4

Dataset can be provided upon request (being a 7.5mb shapefile).

[mmetz]

Replying to dido:

Polygons disappear after v.generalize simplification.

A large rectangular polygon dataset is used with ~14 000 features. Polygons are first processed with rmarea to remove all polygons of size 900 m2. Then a simplification is executed with v.generalize using a factor of 10, all other parameters defaulted.

After simplification several large polygons are missing from the set rendering it useless.

System is a win7x64; QGIS 2.6.1 and GRASS 6.4.4

Dataset can be provided upon request (being a 7.5mb shapefile).

I assume the workflow is v.in.ogr + v.clean tool=rmarea + v.generalize threshold=10.

Several functions relating to v.generalize have just been fixed in trunk and I would like to test. Can you please provide a link to the shapefile or send the shapefile to markus.metz.giswork at gmail dot com? Thanks!

[dido]

Import is done using v.in.ogr with snapping threshold = 1 and min area = 0.000000000001. This one probably engages a v.clean operation also.

I've sent you a temporary download link for the dataset.

[mmetz]

Replying to dido:

Polygons disappear after v.generalize simplification.

A large rectangular polygon dataset is used with ~14 000 features. Polygons are first processed with rmarea to remove all polygons of size 900 m2. Then a simplification is executed with v.generalize using a factor of 10, all other parameters defaulted.

After simplification several large polygons are missing from the set rendering it useless.

I can confirm the problem. This is a bug in the vector libraries regarding topology updating and has been fixed in GRASS 7. Please wait for GRASS 7.0.0RC2 or use a daily build.

[martinl]

Ticket retargeted after milestone closed