+ 0 I'm not sure which approach DesiredFields vs IgnoredFields is the best one. Both have their pros and cons. IgnoredFields will be a bit awkward to use, but the main argument for this approach (avoid dropping any special fields) makes sense.
I also think that a layer capability can be useful to know which driver makes use of it. For Tamas : About the -select option of ogr2ogr, I'd note that, paradoxically ogr2ogr does not use currently the SQL engine to implement this. The list of fields is just used to create the layout of the target layer definition and build the map between source and target field indices to be used by OGRFeature::SetFrom(). At first thought, I have imagined it could generate a "SELECT field1, field2, ... FROM layer_name" SQL expression instead and fallback to the -sql case. But for server based drivers (postgresql or mysql for example), this wouldn't fetch the geometry field... Best regards, Even Le vendredi 06 août 2010 15:04:12, Martin Dobias a écrit : > Hi, > > As there have been no further comments on RFC 29, I would like to call > for vote on it: > > http://trac.osgeo.org/gdal/wiki/rfc29_desired_fields > > Regards > Martin > _______________________________________________ > gdal-dev mailing list > [email protected] > http://lists.osgeo.org/mailman/listinfo/gdal-dev _______________________________________________ gdal-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/gdal-dev
