Peter,

I suspect the reprojection of your features went fine, otherwise you would have got an error message like "Failed to reproject feature XXXX (geometry probably out of source or destination SRS)". This error message must come from some code that tries heuristics related to polar or anti-meridian geometry splitting, although normally there are provisions to silent non-critical reprojection errors related to those heuristics, but perhaps there's some silencing missing somewhere.

I could have a look if you can pass on the dataset (you can send it / a link for it privately if needed)

Even

Le 19/12/2022 à 19:16, Peter Townsend via gdal-dev a écrit :
I've got this test dataset of a bunch of forest areas in Alaska. The shapefile is in Albers, and we're converting it to EPSG:4326 lat longs.

Layer name: trees
Geometry: Polygon
Feature Count: 444
Extent: (-2175230.041280, 1899842.624457) - (4895522.387686, 6909266.538422)
Layer SRS WKT:
PROJCRS["Albers",
    BASEGEOGCRS["NAD27",
        DATUM["North American Datum 1927",
            ELLIPSOID["Clarke 1866",6378206.4,294.978698213898,
                LENGTHUNIT["metre",1]],
            ID["EPSG",6267]],
        PRIMEM["Greenwich",0,
            ANGLEUNIT["Degree",0.0174532925199433]]],
    CONVERSION["unnamed",
        METHOD["Albers Equal Area",
            ID["EPSG",9822]],
        PARAMETER["Latitude of false origin",50,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8821]],
        PARAMETER["Longitude of false origin",-154,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8822]],
        PARAMETER["Latitude of 1st standard parallel",55,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8823]],
        PARAMETER["Latitude of 2nd standard parallel",65,
            ANGLEUNIT["Degree",0.0174532925199433],
            ID["EPSG",8824]],
        PARAMETER["Easting at false origin",0,
            LENGTHUNIT["US survey foot",0.304800609601219],
            ID["EPSG",8826]],
        PARAMETER["Northing at false origin",0,
            LENGTHUNIT["US survey foot",0.304800609601219],
            ID["EPSG",8827]]],
    CS[Cartesian,2],
        AXIS["(E)",east,
            ORDER[1],
            LENGTHUNIT["US survey foot",0.304800609601219,
                ID["EPSG",9003]]],
        AXIS["(N)",north,
            ORDER[2],
            LENGTHUNIT["US survey foot",0.304800609601219,
                ID["EPSG",9003]]]]

We're trying to upgrade our GDAL from 3.21 to 3.53, and our poor test is getting flagged now with either "ERROR 1: Point outside of projection domain" and/or "ERROR 1: tolerance condition error".

This is what I'm seeing:
ogr2ogr.exe -skipfailures -t_srs EPSG:4326 --debug on c:\path\output c:\path\trees.shp GDAL: GDALOpen(c:\path\trees.shp, this=0000027A1189CBE0) succeeds as ESRI Shapefile.
GDAL: Using ESRI Shapefile driver
GDAL: GDALDriver::Create(ESRI Shapefile,c:\path\output,0,0,0,Unknown,0000000000000000)
Shape: DBF Codepage = LDID/87 for c:\path\trees.shp
Shape: Treating as encoding 'ISO-8859-1'.
ERROR 1: tolerance condition error
ERROR 1: tolerance condition error
GDALVectorTranslate: 444 features written in layer 'trees'
Shape: 444 features read on layer 'trees'.
GDAL: GDALClose(c:\path\trees.shp, this=0000027A1189CBE0)
GDAL: GDALClose(c:\path\output, this=0000027A1189D4A0)
GDAL: In GDALDestroy - unloading GDAL shared library.

(Same thing happens without skipfailures.)

Now I'm 95% sure it's throwing this error because there's some vertex somewhere in here that's in that special "dead zone" for Alaska Albers that can't convert to 4326 (seen this before on other datasets trying to project individual points).

What I'm wondering is this:
1) Is there any way to get GDAL to tell me what feature/coordinate threw the error? 2) What is it doing as a fallback behavior? Skipping the point and moving on? I still come out with the same number of features in the end.

Thanks!
--
Peter Townsend
Senior Software Developer

_______________________________________________
gdal-dev mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/gdal-dev

--
http://www.spatialys.com
My software is free, but my time generally not.
_______________________________________________
gdal-dev mailing list
[email protected]
https://lists.osgeo.org/mailman/listinfo/gdal-dev

Reply via email to