JarCastingEASYXML
Easyxml an approach for handling large XML files with low memory footprint. The idea is to use the StAX reader and convert only a specific XML nodes to JDom2 or JAXB (at the moment). With this combination you get a very low memory footprint with the luxury of using high level XML parser.
The project requires Java 8.
|
Note
|
The project is still heavily under development. It would be nice if you have some tips to improve the API. |
Example XML
For the next description of parsing XML files we are using the following XML as an basis for the further examples.
<notes>
<note type="simple">
<id>1</note>
<content>first note</content>
</note>
<note type="simple">
<id>2</note>
<content>second note</content>
</note>
...
<note type="simple">
<id>10000</note>
<content>xth note</content>
</note>
</notes>
JDom2
The first example shows the handling of XML files with JDOM2.
public class Note {
private Long id;
private String content;
}
try(InputStream inputStream = new FileInputStream(new File("note.xml"))){
Parser<Element, Note> parser = JDom2ReaderBuilder
.<Note> reader()
.setInputStream(inputStream)
.addItemReader(
JDom2ItemReaderBuilder
.<Note> itemReader()
.shouldHandle("notes/note")
.handle((e) -> {
Note note = new Note();
note.setId(Long.valueOf(e.getChildTextTrim("id")));
note.setContent(e.getChildTextTrim("content"));
return note;
}).build()
)
.build();
for (Note note : parser) {
System.out.println(note);
}
}
}
The noticeable parts will we explained in the folowing
The shouldHandle method will be called for each XML start element which is found in the document while parsing it with StAX. With the shouldHandle method you can consume the StAX stream to convert it to an JDom2 element. The shouldHandle method expects the path in the DOM tree.
.shouldHandle("notes/note")
When the path in the shouldHandle method matches the current path then the code within the handle method is called. The handle function expects an Java 8 Function<Element,Boolean>. The current element is automatically converted into an JDom2 element and can be accessed as usual.
Inside the handle a new instance of the Note class is created and returnd.
.handle((e) -> {
Note note = new Note();
note.setId(Long.valueOf(e.getChildTextTrim("id")));
note.setContent(e.getChildTextTrim("content"));
return note;
}
for(Note note : parser) {
System.out.println(note);
}
JAXB
The use of JAXB is similar to the JDom2 approach. You have two possibilities to use JAXB
-
via the ObjectFactory which is generated from a XSD
-
via a self generated "Note" Class and an jaxb.index file.
The JaxbReaderBuilder needs the path of the of the ObjectFactory or the jaxb.index file. With the withJaxbClass
try (InputStream inputStream = new FileInputStream(new File("note.xml"))) {
Parser<JAXBElement<Note>, Note> parser = JaxbReaderBuilder
.<Note> reader(ObjectFactory.class.getPackage().getName())
.setInputStream(inputStream)
.addItemReader(
new JaxbItemReaderBuilder<Note>()
.withJaxbClass(Note.class)
.shouldHandle("notes/note")
.build()
)
.build();
for (Note note : parser) {
System.out.println(note);
}
}
Update with JDom2
With the help of the Jdom2WriterBuilder you can modify existing XML files without knowing the hole structure of the document. The approach is the same as with the JDom2Reader, with it you can modify the document within the handle and the changes are automatically written back to the output stream.
We can also remove specific parts of the document with the remove method instead of using the handle method.
try (
InputStream inputStream = new FileInputStream(new File("note.xml"));
OutputStream outputStream = new FileOutputStream(new File("note_modified.xml"))) {
Writer writer = Jdom2WriterBuilder.writer()
.setInputStream(inputStream)
.setOutputStream(outputStream)
.addItemWriter(
new Jdom2ItemWriterBuilder()
.shouldHandle(p -> {
NoteContext noteContext = (NoteContext) p;
return p.getPath().equals("notes/group/note") && noteContext.latestGroupId == 2;
})
.remove()
.build())
.addItemWriter(
new Jdom2ItemWriterBuilder()
.shouldHandle("notes/note")
.handle((c) -> {
Element contentElement = c.getElement().getChild("content");
contentElement.setText("content: " + contentElement.getText());
})
.build())
.build();
writer.writeAll();
}