Frederik Ramm wrote: > There should also be a number of ways referring to nodes that don't exist? Yup, but the table with way_nds is still to be inserted correctly. One big memory issue on a 2GB of memory system :( If I manage I'll copy it too. Stefan