[gdal-dev] Missing features after copying layers from Shapefiles to SQLite

[Nik Sands]

Thanks to everybody who helped me towards getting this working - it is now working correctly.  The conversion process is a little slow for large data sets, but the results are excellent, and the Spatialite databases are MUCH faster for spatial filtered lookups than the Shapefiles I had been using before.

In the end it turns out that I had my code very nearly correct before looking at the code in ogr2ogr from which I then corrected my previous cludgy force-to-multi code to the following:

	OGR_F_SetGeometryDirectly(feature, OGR_G_ForceToMultiPoint(OGR_F_StealGeometry(feature)));

(and other variants depending on the geometry type, of course).


On 4 Mar 2014, at 9:58 pm, Nik Sands <nixanz at nixanz.com> wrote:

> Thanks for this information.
> 
> I'd like to replicate that ogr2ogr PROMOTE_TO_MULTI functionality in my code, but I'm having trouble figuring out how to do it.  I guess I can take a look at the ogr2ogr code.
> 
> (I could also just include the entire ogr2ogr tool bundled into my app and call it from the app, but that seems a bit of overkill for an iOS app, which is best as lean as possible.)
> 
> 
> On 4 Mar 2014, at 9:50 pm, a.furieri at lqt.it wrote:
> 
>> Hi all,
>> 
>> just yo better clarify some relevant details:
>> 
>> a Shapefile can legitimately contain one the following shapes:
>> Point, PolyLine, Polygon and MultiPoint; and there are further
>> shape-codes intended to support extended XYZ or XYM dimensions,
>> such as PointZ and PolyLineM
>> 
>> please note: accordingly to the SHP model there is a very strong
>> distinction between Point and MultiPoint; but a PolyLine can
>> indifferently contain just a single line or a collection of
>> many individual lines, and the same is for Polygons.
>> 
>> on the opposite side, the standard OGC-SFS data model supports
>> the following geometry classes:
>> Point, Linestring, Polygon, MultiPoint, MultiLinestring,
>> MultiPolygon and GeometryCollection; in this case too we have
>> further classes supporting extending XYZ, XYM and XYZM
>> dimensions, such as PointZ, LinestringM or PolygonZM.
>> 
>> please note: the OGC-SFS model introduces a very sharp distinction
>> between e.g. Linestring and MultiLinestring; a Linestring can
>> contain just a single line, a MultiLinestring can contain any
>> arbitrary number of individual lines ... THIS INCLUDING JUST ONE !!!
>> and the same is for Point/MultiPoint and Polygon/MultiPolygon.
>> 
>> quick conclusion: SHP geometry types doesn't exactly match
>> OFC-SFS types, and there is enough room to introduce some
>> dangerous inconsistency, e.g.:
>> 
>> - the SHP PolyLine type could possibly match the Linestring type
>> if (and only if) all PolyLines into the SHP will contain just
>> a single line.
>> - when some SHP PolyLine does actually contain more individual
>> lines a MultiLinestring will always be required on the OGC-SFS
>> side
>> - anyway OGC-SFS MultiLinestring will always exactly match the
>> SHP PolyLine type, even in the case of "mixed" Polylines
>> sometimes being elementary and sometimes being complex.
>> - the same considerations obviously apply to Polygon/MultiPolygon
>> - conclusion: always assuming a Multi?? type is the safest approach,
>> if you aren't absolutely sure about the actual SHP payload.
>> 
>> ogr2ogr does actually has an argument nicely supporting
>> all this (just pasted from the ogr2ogr man page):
>> 
>> -nlt type:
>>   Define the geometry type for the created layer.
>>   One of NONE, GEOMETRY, POINT, LINESTRING, POLYGON,
>>   GEOMETRYCOLLECTION, MULTIPOINT, MULTIPOLYGON or
>>   MULTILINESTRING. Add "25D" to the name to get 2.5D
>>   versions. Starting with GDAL 1.10, PROMOTE_TO_MULTI
>>   can be used to automatically promote layers that mix
>>   polygon or multipolygons to multipolygons, and layers
>>   that mix linestrings or multilinestrings to multilinestrings.
>>   Can be usefull when converting shapefiles to PostGIS
>>   (and other target drivers) that implements strict checks
>>   for geometry type.
>> 
>> it works for PostGIS, and it works for SpatiaLite as well.
>> 
>> bye Sandro
>> 
>> _______________________________________________
>> gdal-dev mailing list
>> gdal-dev at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/gdal-dev
> 
> _______________________________________________
> gdal-dev mailing list
> gdal-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/gdal-dev