Marshalling XML with large base64 content

Question

i have a XML-Schema which contains base64 binary data. The problem is, if the binary is big enough I unsurprisingly get an OutOfMemoryError. I managed to generate the affected java class to use the DataHanlder instead of a byte[] but still JAXb seams to do the marshaling in RAM. The used schema cannot be changed and is very complex so building the XML by handy is not a solution. The only idea i have on this is adding a placeholder instead of the big binary and replacing it afterwards. But I believe there is a better solution!?

Thanks for your hints

Sample Schema:

Generated Java Class:

package com.example.gen;

import javax.activation.DataHandler;
import javax.xml.bind.annotation.XmlAccessType;
import javax.xml.bind.annotation.XmlAccessorType;
import javax.xml.bind.annotation.XmlElement;
import javax.xml.bind.annotation.XmlMimeType;
import javax.xml.bind.annotation.XmlRootElement;
import javax.xml.bind.annotation.XmlType;

@XmlAccessorType(XmlAccessType.FIELD)
@XmlType(name = "", propOrder = {
    "text"
})
@XmlRootElement(name = "Document")
public class Document {
    @XmlElement(required = true)
    @XmlMimeType("anything/else")
    protected DataHandler text;
    public DataHandler getText() {
        return text;
    }
    public void setText(DataHandler value) {
        this.text = value;
    }
}

Sample Code:

File bigFile = new File("./temp/bigFile.bin");
File outFile = new File("./temp/bigXML.xml");

Document document = new Document();
DataHandler bigDocDH = new DataHandler(new FileDataSource(bigFile));
document.setText(bigDocDH);

JAXBContext jaxbContext = JAXBContext.newInstance("com.example.gen");
Marshaller marshaller = jaxbContext.createMarshaller();

OutputStream outputStream = new FileOutputStream(outFile); 
marshaller.marshal(document, outputStream);

Mani · Accepted Answer

Ok, i have found a solution which works for me: First i replace the DataHandler which points to a large file with a DataHandler which contains a small byte array as content.

After this i implemented a XMLStreamWriterWrapper which delegates all methods to another XMLStreamWriter. If the content of the Datahandler with simple content is written to the XMLSteamWriterWrapper i remove the data and stream the original data to this position.

Constructor and Factory:

/**
 * Constructor.
 * 
 * @param outputStream
 *            {@link #outputStream}
 * @param binaryData
 *            {@link #binaryData}
 * @param token
 *            the search token.
 * @throws XMLStreamException
 *             In case the XMLStreamWriter cannot be constructed.
 */
private XMLStreamWriterWrapper(OutputStream outputStream, DataHandler binaryData, String token) throws XMLStreamException {
    this.xmlStreamWriter = XMLOutputFactory.newFactory().createXMLStreamWriter(outputStream);

    // ensure the OutputStream is buffered. otherwise encoding of large data
    // takes hours.
    if (outputStream instanceof BufferedOutputStream) {
        this.outputStream = outputStream;
    } else {
        this.outputStream = new BufferedOutputStream(outputStream);
    }
    this.binaryData = binaryData;
    // calculate the token.
    byte[] encode = Base64.getEncoder().encode(token.getBytes(Charset.forName("UTF-8")));
    this.tokenAsString = new String(encode, Charset.forName("UTF-8"));
    this.token = this.tokenAsString.toCharArray();
}

/**
 * Factory method to create the {@link XMLStreamWriterWrapper}.
 * 
 * @param outputStream
 *            The OutputStream where to marshal the xml to.
 * @param binaryData
 *            The binary data which shall be streamed to the xml.
 * @param token
 *            The token which akts as placeholder for the binary data.
 * @return The {@link XMLStreamWriterWrapper}
 * @throws XMLStreamException
 *             In case the XMLStreamWriter could not be constructed.
 */
public static XMLStreamWriterWrapper newInstance(OutputStream outputStream, DataHandler binaryData, String token) throws XMLStreamException {
    return new XMLStreamWriterWrapper(outputStream, binaryData, token);
}

writeCharacters implementation:

/*
 * (non-Javadoc)
 * 
 * @see javax.xml.stream.XMLStreamWriter#writeCharacters(java.lang.String)
 */
@Override
public void writeCharacters(String text) throws XMLStreamException {
    if (this.tokenAsString.equals(text)) {
        writeCharacters(text.toCharArray(), 0, text.length());
    } else {
        xmlStreamWriter.writeCharacters(text);
    }
}

/*
 * (non-Javadoc)
 * 
 * @see javax.xml.stream.XMLStreamWriter#writeCharacters(char[], int, int)
 */
@Override
public void writeCharacters(char[] text, int start, int len) throws XMLStreamException {
    char[] range = Arrays.copyOfRange(text, 0, len);
    if (Arrays.equals(range, token)) {
        LOGGER.debug("Found replace token. Start streaming binary data.");
        // force the XMLStreamWriter to close the start tag.
        xmlStreamWriter.writeCharacters("");
        try {
            // flush the content of the streams.
            xmlStreamWriter.flush();
            outputStream.flush();
            // do base64 encoding.
            OutputStream wrap = Base64.getMimeEncoder().wrap(outputStream);
            this.binaryData.writeTo(wrap);
        } catch (IOException e) {
            throw new XMLStreamException(e);
        } finally {
            try {
                // flush the output stream
                outputStream.flush();
            } catch (IOException e) {
                throw new XMLStreamException(e);
            }
        }
        LOGGER.debug("Successfully inserted binary data.");
    } else {
        xmlStreamWriter.writeCharacters(text, start, len);
    }
}

Example usage:

//Original file DataHandler
DataHandler bigDocDH = new DataHandler(new FileDataSource(bigFile));

Document document = new Document();
String replaceToken = UUID.randomUUID().toString();
//DataHandler with content replaced by the XMLStreamWriterWrapper
DataHandler tokenDH = new DataHandler(new ByteArrayDataSource(replaceToken.getBytes(Charset.forName("UTF-8")), bigDocDH.getContentType()));
document.setText(tokenDH);

try (OutputStream outStream = new FileOutputStream(outFile)) {
    XMLStreamWriter streamWriter = XMLStreamWriterWrapper.newInstance(outStream, bigDocDH, replaceToken);
    marshaller.marshal(document, streamWriter);
}

Marshalling XML with large base64 content

Answers (1)

Related Questions