When I need to convert CAD data into GIS format, the first thing I do is run a topological check. This is important because it helps me identify open polylines that would certainly generate errors during polygon conversion, but it is not always enough.
Sometimes this initial step is still not sufficient, because the classic self-ring intersection appears and in turn generates polygons with infinitesimal areas. These very small polygons sometimes slip past the topological check after converting polylines into polygons, so we think the dataset is clean. We realise that it is not only when we try to perform geoprocessing or selection-by-location operations.
For this reason I follow the second topological check with the calculation of the areas and eliminate all those polygons with infinitesimal areas.
So when I need to extract polygons, my workflow is the following:
- topological check and error cleaning
- converting lines to polygons
- topological check and error cleaning
- surface calculation and polygon removal with infinitesimal surfaces