All the data is there, you could run it from the beginning and rebuild the whole thing. There is less than 5 million rows to parse in the nplb database.