Well it's been a while. I've had muchos problems with
getting the NaPTAN & TNDS dataset into a database within a reasonable amount of time. We're talking 10,000+ .xml files and millions of database rows so the method of importing makes it the difference between minutes and days. After trying Pentaho data integration and a few others options I've settled with coding the import script in node.js using a SAX parser.
As for Node.js I chose it as it provides good support for asynchronous operations, good performance, some decent libraries, and it's great for the more web app focused sites.