mirror of
https://github.com/IfcOpenShell/IfcOpenShell.git
synced 2026-08-09 17:31:45 +00:00
fee6ea52ba
f attributes holding entity lists are expanded in SQL, this means that you end up with significantly more rows. Also, you cannot have the ifc_id marked as a UNIQUE field / benefit from unique primary key index. So leaving the fields as JSON is actually significantly more efficient (almost 10X faster to load a model into BBIM).