Disclaimer: I am everything but a Blender expert.
When I go to https://www.shapenet.org/taxonomy-viewer to the car category and click on a model and click on Source I end up on 3dwarehouse.sketchup website where I can download a collada file (readable by blender).
For instance https://3dwarehouse.sketchup.com/model.html?id=2d096d5dc0e0eeb370a43c2d978e502e shows a 2006 Honda Civic.
After importing it in Blender (File -> Import -> Collada) I end up with different results depending on the model I clicked on:
-I cannot see anything except the two extreme points
-I can see the body of the car and wheels in two different places
-I can see the body of the car without the tires
-I can see the whole thing normally
In every case whenever I render it I also get black texture artifacts everywhere.
However in shapenet viewer everything is perfectly normal in every case !
How to import it correctly to blender to render it and obtain approximately the same kind of render shapenet displays ?