Yes, Bernhard is right. Actually, this might be caused by rounding issues (or by the limit of floating point calculations).
I often have problems with topology when working with another projection. The direct and its invers calculation hardly is the same. Am Mi, 31.05.2017, 10:53 schrieb Bernhard Ströbl: > Andreas, > when loading your shape file into a new project the project is set to > EPSG:4326 because your data are in EPSG:4326 > Your screenshots indicate that you are working in another projection. In > EPSG:4326 split parts works flawlessly, when I use a projection I get > the invalid geometries error. > In my experience you should edit geometries in a project matching the > projection of the data, although editing projected geometries is not > prohibited by QGIS. > > Bernhard > > Am 31.05.2017 um 09:49 schrieb Andreas Wicht: >> Hi James, >> hi Bernhard, >> >> attached you can find a small sample data set and three screenshots. >> I used the Split Parts tool to cut off the lower part as shown in the >> screenshots. >> >> This is in a PostGIS layer >> (POSTGIS="2.3.2 r15302" GEOS="3.5.0-CAPI-1.9.0 r4084" PROJ="Rel. >> 4.9.2, 08 September 2015" GDAL="GDAL 1.11.3, released 2015/09/16" >> LIBXML="2.9.3" LIBJSON="0.11.99" TOPOLOGY RASTER) >> >> geometrytype >> -------------- >> MULTIPOLYGON >> >> I tested with QGIS 2.18.9 in Windows 7 and Ubuntu 16.04 (ubuntugis >> deps). >> ST_IsValidReason yields the following after editing (depicted in third >> screenshot): >> >> NOTICE: Self-intersection at or near point -19.203503991278154 >> 65.808265481598852 >> st_isvalidreason >> ------------------------------------------------------- >> Self-intersection[-19.2035039912782 65.8082654815989] >> >> I can reproduce this behaviour with simple hand drawn polygons in a >> memory layer as well. >> My described workflow in the earlier email was aiming at the comment >> (#16) of Jürgen Fischer in the mentioned issue >> (https://issues.qgis.org/issues/12799#note-16). Here he describes that >> further editing of the geometry is essential to use this function >> (please correct me, if I understood that wrong). >> >> By current status of development Bernhard's approach seems to be the >> most efficient one. >> >> Yet I think that the function should be designed in a way that it does >> not require further validation/editing to get a clean geometry. That >> is not obvious to a less experienced user. >> >> cheers >> Andreas >> >> >> On 31 May 2017 at 02:47, James Wood <[email protected]> wrote: >>> Hey Andreas, >>> I haven't experienced the invalid geometry as you describe, but I did >>> find that my QGIS install had some issues. I have upgraded to 2.18.9 on >>> Win10. If you have a sample of your data available, I would be happy to >>> try to corroborate. >>> >>> James >>>> On May 29, 2017, at 06:16, Andreas Wicht <[email protected]> wrote: >>>> >>>>> On 28 May 2017 at 14:43, James Wood <[email protected]> wrote: >>>>> On the digitizing toolbar, try "Split Parts" instead of "Split >>>>> Features" on multipart polygons. >>>> >>>> Thank you for the hint, James. Somehow I always thought that this >>>> function is only suitable for separating an existing part from a >>>> feature (never used it though). >>>> Yet, if you cut a polygon with the "Split Parts" tool, the output will >>>> be an invalid geometry (I observed self-intersections). This as well >>>> can not be the intended behavior. I could observe those errors in >>>> simple scratch layers as well as in PostGIS layers. >>>> >>>> Has anybody else experienced this behavior? >> >> >> >> __________ Information from ESET Mail Security, version of virus >> signature database 15504 (20170531) __________ >> >> The message was checked by ESET Mail Security. >> http://www.eset.com >> > > > > __________ Information from ESET Mail Security, version of virus signature > database 15505 (20170531) __________ > > The message was checked by ESET Mail Security. > http://www.eset.com > > > _______________________________________________ > Qgis-user mailing list > [email protected] > List info: https://lists.osgeo.org/mailman/listinfo/qgis-user > Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user _______________________________________________ Qgis-user mailing list [email protected] List info: https://lists.osgeo.org/mailman/listinfo/qgis-user Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
