Fastest and best way to transform large XML docs from one format to another [closed]

问题

I have a bunch of large XML files (total size of all files is more than 1 GB) and I need to transform them from a vendor schema to our schema.

The vendor has one ZIP file (it contains large XML files) at some FTP location on its server. I have to pick that ZIP file up and then transform all available XML files. After transforming to our schema format, I need to persist the data in a database.

What is a good design to implement this? What are relevant tools and utilities which support Java?

回答1:

Just use the regular Java API...

File zipFile = new File("archive.zip");
File xsltFile = new File("transform.xslt");
File transformedXmlsFile = new File("transformed.xml");

StreamSource xsltSource = new StreamSource(xsltFile);
Transformer transformer = TransformerFactory.newInstance().newTransformer(xsltSource);

ZipInputStream zipIn = new ZipInputStream(new FileInputStream(zipFile));
ZipEntry zipEntry;
OutputStream resultXmls = new FileOutputStream(transformedXmlsFile);
while ((zipEntry = zipIn.getNextEntry()) != null){
  StreamSource inputXml = new StreamSource(zipIn);
  StreamResult resultXml = new StreamResult(resultXmls);
  transformer.transform(inputXml, resultXml);
}
zipIn.close();
resultXmls.close();

回答2:

I like the simple methods. I would use any SAX,Stax implementation, and do not use any DOM. But it just me, maybe you will find here a fancy library, which will do all work instead of you :)

1GB xml in DOM can eat your all ram! - be carefully what library are you choosing and what he use behind of scene.

I hope it helps!

回答3:

I used Saxon-EE for transformation and Woodstox for XML unmarshalling.

来源：https://stackoverflow.com/questions/12482729/fastest-and-best-way-to-transform-large-xml-docs-from-one-format-to-another

标签

java

design-patterns

architecture

implementation