Doc: gdal vector check-geometry #13844
Description
Feature description
I am going to add some examples to the https://gdal.org/en/latest/programs/gdal_vector_check_geometry.html page, and was first testing the tool.
This issue contains some initial thoughts and suggestions on possible improvements - no bugs were found, and it is a very useful tool.
cc @dbaston
- I initially wanted to see if there were any geometry errors in a layer in an ESRI FileGeoDatabase:
gdal vector check-geometry C:\Data\test.gdb.zip --input-layer "TestLayer" --output-format CSV /vsistdout/
error
At first, I thought this was a nondescript error message, until running further tests below when I realised it was the field name of the output and there were no errors in the dataset. Maybe the field could be renamed errorDesc or similar? I will add a note on the default output fields to the docs page.
- Testing on multiple layers in a dataset:
gdal vector check-geometry C:\Data\test.gdb.zip C:\Data\errors.gpkg
In my test dataset there were 40 layers. The geopackage is created with 40 output layers, however the majority of these were empty. It takes a few minutes to manually find which layers have errors in an application such as QGIS.
An --ignore-valid flag would be very useful to avoid empty output. Alternatively I can add a pipeline example that filters these out.
Being able to filter output would also be useful on the command-line, to avoid noise:
gdal vector check-geometry C:\Data\test.gdb.zip --quiet -f CSV --lco GEOMETRY=AS_XY --lco SEPARATOR=TAB /vsistdout/
X Y error
self-intersection
X Y error
X Y error
X Y error
X Y error
X Y error
self-intersection
X Y error
self-intersection
X Y error
self-intersection
X Y error
X Y error
...
- Allowing all fields to be easily added to the output would be very handy, using the same keywords as https://gdal.org/en/latest/programs/gdal_raster_pixel_info.html -
ALLandNONE(I'm not sure what happens here if you have a field namedALLin the dataset).
--include-field <INCLUDE-FIELD> Fields from coordinate dataset to include in output (special values: ALL and NONE) (default: ALL) [may be repeated]
- This may be related to my test data, but there is a
MULTIPOINTwarning in the tool output:
gdal vector check-geometry C:\Data\test.gdb.zip C:\Data\errors.gpkg
Warning 1: Point geometry passed to 'gdal vector check-geometry'. Point geometries are always valid/simple. Further messages of this type will be suppressed.
Warning 1: A geometry of type MULTIPOINT is inserted into layer error_location_TestLayer of geometry type POINT, which is not normally allowed by the GeoPackage specification,
but the driver will however do it. To create a conformant GeoPackage, if using ogr2ogr, the -nlt option can be used to override the layer geometry type. This warning will no longer be emitted for this combination of layer and feature geometry type.
- The current docs state Validity/simplicity checking is performed by the GEOS library - is there a list of possible errors to add to the docs page, or to link to? I found a few of the errors I encountered in this enum, but I presume there is a list / details somewhere we can link to.
Additional context
No response