Nested holes when using ogr2ogr mid-mif -> postgis

[poulet1212]

When using ogr2ogr tool to convert multipolygons having nested holes from mid-mif to postgis, the resulting postgis table does not contain nested hole. In fact, these holes exists but they are "filled" with a polygon.

Ive seen a post about it there: ​http://medspx.homelinux.org/blog/Qgis/postgis_nested_holes.html (sorry its in french)

The weird is that if I convert a SHAPE dataset that contains nested holes to postgis, these are converted correctly.

-Ignore the first file. The zip files contain vegetal coverage polygons, it contains nested holes that are buildings.

The cammand I used is like:

ogr2ogr -progress -a_srs EPSG:21781 -f PostgreSQL PG:"dbname=postgis host=localhost port=5432 user=postgres password=postgis" -nln mid-mif-exported-table C:\temp\MOVD_CAD_TPR_CSVERT_S.mif MOVD_CAD_TPR_CSVERT_S

I can attach the mid/mif and shp files I used.

[Even Rouault]

Not sure if this is really an issue with the MIF/MID driver or with the data file. When doing a ogrinfo on it, I can see inner rings being reported for some polygons.

[poulet1212]

Ok. Do you have a sample mid/mif dataset with nested holes that I can check if they appear correctly once transformed in postgis?

[Daniel Morissette]

I see the version is set to 1.7.2 in the ticket. Could this be something that was fixed in 1.8.0? That would explain why rouault sees the inner rings and poulet1212 doesn't.

I know we've added calls to OGRGeometryFactory::organizePolygons for mif features at some point, but I'm not sure to which GDAL/OGR version that corresponds.

[poulet1212]

Ok, I've tried with the latest 1.8 build and now the inner rings appear correctly when transforming mid/mif files to Postgis. So it is something that has been fixed in 1.8 and still I use 1.7 I did not see it. However I'm gonna move to this version and also follow the new ones :-)

Thanks a lot for your comments.