Bug report #2089

Buffer produces incorrect results with multipart vectors

Added by jaanvajakas - almost 14 years ago. Updated almost 13 years ago.

Assignee:cfarmer -
Category:Python plugins
Affected QGIS version: Regression?:No
Operating System:All Easy fix?:No
Pull Request or Patch supplied: Resolution:invalid
Crashes QGIS or corrupts data: Copied to github as #:12149


To reproduce:
1. create a new project in QGIS,
2. add layer kirdem2.epg4326.shp using Layers -> Add vector layer...,
3. Tools -> Geoprocessing tools -> Buffer(s), set buffer distance to -0.0003, choose output file, click OK

Expected result: an inner buffer of the polygons in platte carree distance (not real distance, as interpreting degrees as meters would be ambiguous)

Actual result: certainly not a buffer neither in platte carree distance nor in real geodetic distance nor any other projection. In some places, the buffer exceeds the original polygon border (see qgis_buffer_bug1.png); in some places, the buffer has "pixel artifacts" (qgis_buffer_bug1.png); some islands are missing altogether (qgis_buffer_bug2.png).

The input data kirdem2.epg4326.shp is part of the publicly available dataset '01.09.2009a. asustus├╝ksus SHP' (zip - 13,65 MB, 6.11.2009 ) on
(I cut a small part out and converted to EPSG:4326 using QGIS).

kirdem2.epsg4326.zip - the input dataset (152 KB) jaanvajakas -, 2009-11-13 07:51 AM

qgis_buffer_bug1.png (113 KB) jaanvajakas -, 2009-11-13 07:56 AM

qgis_buffer_bug2.png (120 KB) jaanvajakas -, 2009-11-13 07:57 AM


#1 Updated by Giovanni Manghi almost 14 years ago

If you transform your vector from multipart to singleparts the buffer results ok.

This may be a simple workaround or it may be simply necessary, I'm not sure how buffer tools is supposed to work with multipart vectors.


#2 Updated by Giovanni Manghi over 13 years ago

awaiting Carson reply, I'll push this to 1.6 as in any case there is a workaround

#3 Updated by cfarmer - almost 13 years ago

  • Resolution set to invalid
  • Status changed from Open to Closed

Tested using f13d3f35 (SVN r14368).

I am unable to reproduce this error. Looking at the areas described in this ticket, results appear to be as expected. I suspect this was a GEOS error, but appears to now have been fixed (GEOS version?). Please test and confirm if this now works.

I am going to close this as "invalid" because this is now a relatively old ticket, and appears to now be invalid, however, if it is still a problem, please reopen this ticket.


Also available in: Atom PDF