LOGLINK table is processed after all other records have been processed, regardless of their order in the import data stream. This is because the validity of the series depends on the existence and type of connections, some of which may not be mentioned until later in the import file.