如何提高分割xml文件的性能

Question

I've see quite a lot posts/blogs/articles about splitting XML file into a smaller chunks and decided to create my own because I have some custom requirements. Here is what I mean, consider the following XML :

<?xml version="1.0" encoding="UTF-8" standalone="no" ?> 
<company>
 <staff id="1">
    <firstname>yong</firstname>
    <lastname>mook kim</lastname>
    <nickname>mkyong</nickname>
    <salary>100000</salary>
   </staff>
 <staff id="2">
    <firstname>yong</firstname>
    <lastname>mook kim</lastname>
    <nickname>mkyong</nickname>
    <salary>100000</salary>
   </staff>
 <staff id="3">
    <firstname>yong</firstname>
    <lastname>mook kim</lastname>
    <nickname>mkyong</nickname>
    <salary>100000</salary>
   </staff>
 <staff id="4">
    <firstname>yong</firstname>
    <lastname>mook kim</lastname>
    <nickname>mkyong</nickname>
    <salary>100000</salary>
   </staff>
 <staff id="5">
    <firstname>yong</firstname>
    <lastname>mook kim</lastname>
    <salary>100000</salary>
   </staff>
</company>

I want to split this xml into n parts, each containing 1 file, but the staff element must contain nickname , if it's not there I don't want it. So this should produce 4 xml splits, each containing staff id starting at 1 until 4.

Here is my code :

public int split() throws Exception{
        BufferedReader br = new BufferedReader(new InputStreamReader(new FileInputStream(inputFilePath)));

        String line;
        List<String> tempList = null;

        while((line=br.readLine())!=null){
            if(line.contains("<?xml version=\"1.0\"") || line.contains("<" + rootElement + ">") || line.contains("</" + rootElement + ">")){
                continue;
            }

            if(line.contains("<"+ element +">")){
                tempList = new ArrayList<String>();
            }
            tempList.add(line);

            if(line.contains("</"+ element +">")){
                if(hasConditions(tempList)){
                    writeToSplitFile(tempList);
                    writtenObjectCounter++;
                    totalCounter++;
                }
            }

            if(writtenObjectCounter == itemsPerFile){
                writtenObjectCounter = 0;
                fileCounter++;          
                tempList.clear();
            }
        }

        if(tempList.size() != 0){
        writeClosingRootElement();
        }

        return totalCounter;
    }

    private void writeToSplitFile(List<String> itemList) throws Exception{
        BufferedWriter wr = new BufferedWriter(new FileWriter(outputDirectory + File.separator + "split_" + fileCounter + ".xml", true));
        if(writtenObjectCounter == 0){
        wr.write("<" + rootElement + ">");
        wr.write("\n");
        }

        for (String string : itemList) {
            wr.write(string);
            wr.write("\n");
        }

        if(writtenObjectCounter == itemsPerFile-1)
        wr.write("</" + rootElement + ">");
        wr.close();
    }

    private void writeClosingRootElement() throws Exception{
        BufferedWriter wr = new BufferedWriter(new FileWriter(outputDirectory + File.separator + "split_" + fileCounter + ".xml", true));
        wr.write("</" + rootElement + ">");
        wr.close();
    }

    private boolean hasConditions(List<String> list){
        int matchList = 0;

        for (String condition : conditionList) {
            for (String string : list) {
                if(string.contains(condition)){
                    matchList++;
                }
            }
        }

        if(matchList >= conditionList.size()){
            return true;
        }

        return false;
    }

I know that opening/closing stream for each written staff element which does impact the performance. But if I write once per file(which may contain n number of staff ). Naturally root and split elements are configurable.

Any ideas how can I improve the performance/logic? I'd prefer some code, but good advice can be better sometimes

Edit:

This XML example is actually a dummy example, the real XML which I'm trying to split is about 300-500 different elements under split element all appearing at the random order and number varies. Stax may not be the best solution after all?

Bounty update :

I'm looking for a solution(code) that will:

• Be able to split XML file into n parts with x split elements(from the dummy XML example staff is the split element).

• The content of the spitted files should be wrapped in the root element from the original file(like in the dummy example company)

• I'd like to be able to specify condition that must be in the split element ie I want only staff which have nickname, I want to discard those without nicknames. But be able to also split without conditions while running split without conditions.

• The code doesn't necessarily have to improve my solution(lacking good logic and performance), but it works.

And not happy with "but it works". And I can't find enough examples of Stax for these kind of operations, user community is not great as well. It doesn't have to be Stax solution as well.

I'm probably asking too much, but I'm here to learn stuff, giving good bounty for the solution I think.

Answer 1

First piece of advice: don't try to write your own XML handling code . Use an XML parser - it's going to be much more reliable and quite possibly faster.

If you use an XML pull parser (eg StAX ) you should be able to read an element at a time and write it out to disk, never reading the whole document in one go.

Answer 2

Here's my suggestion. It requires a streaming XSLT 3.0 processor: which means in practice that it needs Saxon-EE 9.3.

<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="3.0">

<xsl:mode streamable="yes">

<xsl:template match="/">
  <xsl:apply-templates select="company/staff"/>
</xsl:template>

<xsl:template match=staff">
  <xsl:variable name="v" as="element(staff)">
    <xsl:copy-of select="."/>
  </xsl:variable>
  <xsl:if test="$v/nickname">
    <xsl:result-document href="{@id}.xml">
      <xsl:copy-of select="$v"/>
    </xsl:result-document>
  </xsl:if>
</xsl:template>

</xsl:stylesheet>

In practice, though, unless you have hundreds of megabytes of data, I suspect a non-streaming solution will be quite fast enough, and probably faster than your hand-written Java code, given that your Java code is nothing to get excited about. At any rate, give an XSLT solution a try before you write reams of low-level Java. It's a routine problem, after all.

Answer 3

You could do the following with StAX:

Algorithm

1. Read and hold onto the root element event.
2. Read first chunk of XML:
   1. Queue events until condition has been met.
   2. If condition has been met:
      1. Write start document event.
      2. Write out root start element event
      3. Write out split start element event
      4. Write out queued events
      5. Write out remaining events for this section.
   3. If condition was not met then do nothing.
3. Repeat step 2 with next chunk of XML

Code for Your Use Case

The following code uses StAX APIs to break up the document as outlined in your question:

package forum7408938;

import java.io.*;
import java.util.*;

import javax.xml.namespace.QName;
import javax.xml.stream.*;
import javax.xml.stream.events.*;

public class Demo {

    public static void main(String[] args) throws Exception  {
        Demo demo = new Demo();
        demo.split("src/forum7408938/input.xml", "nickname");
        //demo.split("src/forum7408938/input.xml", null);
    }

    private void split(String xmlResource, String condition) throws Exception {
        XMLEventFactory xef = XMLEventFactory.newFactory();
        XMLInputFactory xif = XMLInputFactory.newInstance();
        XMLEventReader xer = xif.createXMLEventReader(new FileReader(xmlResource));
        StartElement rootStartElement = xer.nextTag().asStartElement(); // Advance to statements element
        StartDocument startDocument = xef.createStartDocument();
        EndDocument endDocument = xef.createEndDocument();

        XMLOutputFactory xof = XMLOutputFactory.newFactory();
        while(xer.hasNext() && !xer.peek().isEndDocument()) {
            boolean metCondition;
            XMLEvent xmlEvent = xer.nextTag();
            if(!xmlEvent.isStartElement()) {
                break;
            }
            // BOUNTY CRITERIA
            // Be able to split XML file into n parts with x split elements(from
            // the dummy XML example staff is the split element).
            StartElement breakStartElement = xmlEvent.asStartElement();
            List<XMLEvent> cachedXMLEvents = new ArrayList<XMLEvent>();

            // BOUNTY CRITERIA
            // I'd like to be able to specify condition that must be in the 
            // split element i.e. I want only staff which have nickname, I want 
            // to discard those without nicknames. But be able to also split 
            // without conditions while running split without conditions.
            if(null == condition) {
                cachedXMLEvents.add(breakStartElement);
                metCondition = true;
            } else {
                cachedXMLEvents.add(breakStartElement);
                xmlEvent = xer.nextEvent();
                metCondition = false;
                while(!(xmlEvent.isEndElement() && xmlEvent.asEndElement().getName().equals(breakStartElement.getName()))) {
                    cachedXMLEvents.add(xmlEvent);
                    if(xmlEvent.isStartElement() && xmlEvent.asStartElement().getName().getLocalPart().equals(condition)) {
                        metCondition = true;
                        break;
                    }
                    xmlEvent = xer.nextEvent();
                }
            }

            if(metCondition) {
                // Create a file for the fragment, the name is derived from the value of the id attribute
                FileWriter fileWriter = null;
                fileWriter = new FileWriter("src/forum7408938/" + breakStartElement.getAttributeByName(new QName("id")).getValue() + ".xml");

                // A StAX XMLEventWriter will be used to write the XML fragment
                XMLEventWriter xew = xof.createXMLEventWriter(fileWriter);
                xew.add(startDocument);

                // BOUNTY CRITERIA
                // The content of the spitted files should be wrapped in the 
                // root element from the original file(like in the dummy example
                // company)
                xew.add(rootStartElement);

                // Write the XMLEvents that were cached while when we were
                // checking the fragment to see if it matched our criteria.
                for(XMLEvent cachedEvent : cachedXMLEvents) {
                    xew.add(cachedEvent);
                }

                // Write the XMLEvents that we still need to parse from this
                // fragment
                xmlEvent = xer.nextEvent();
                while(xer.hasNext() && !(xmlEvent.isEndElement() && xmlEvent.asEndElement().getName().equals(breakStartElement.getName()))) {
                    xew.add(xmlEvent);
                    xmlEvent = xer.nextEvent();
                }
                xew.add(xmlEvent);

                // Close everything we opened
                xew.add(xef.createEndElement(rootStartElement.getName(), null));
                xew.add(endDocument);
                fileWriter.close();
            }
        }
    }

}

Answer 4

@Jon Skeet is spot on as usual in his advice. @Blaise Doughan gave you a very basic picture of using StAX (which would be my preferred choice, although you can do basically the same thing with SAX). You seem to be looking for something more explicit, so here's some pseudo code to get you started (based on StAX):

1. find first "staff" StartElement
2. set a flag indicating you are in a "staff" element and start tracking the depth (StartElement is +1, EndElement is -1)
3. now, process the "staff" sub-elements, grab any of the data you care about and put it in a file (or where ever)
4. keep processing until your depth reaches 0 (when you find the matching "staff" EndElement)
5. unset the flag indicating you are in a "staff" element
6. search for the next "staff" StartElement
7. if found, go to 2. and repeat
8. if not found, document is complete

EDIT:

wow, i have to say i'm amazed at the number of people willing to do someone else's work for them. i didn't realize SO was basically a free version of rent-a-coder.

Answer 5

@Gandalf StormCrow: Let me divide your problem into three separate issues:- i) Reading XML and simultaenous split XML in best possible way

ii) Checking condition in split file

iii) If condition met, process that spilt file.

for i), there are ofcourse mutliple solutions: SAX, STAX and other parsers and as simple as that as you mentioned just read using simple java io operations and search for tags.

I believe SAX/STAX/simple java IO, anything will do. I have taken your example as base for my solution.

ii) Checking condition in split file: you have used contains() method to check for existence of nickname. This does not seem best way: what if your conditions are as complex as if nickname should be present but length>5 or salary should be numeric etc.

I would use new java XML validation framework for this which make uses of XML schema.Please note we can cache schema object in memory so to reuse it again and again. This new validation framework is pretty fast.

iii) If condition met, process that spilt file. You may want use java concurrent APIs to submit async tasks(ExecutorService class) to acheive parallel execution for faster performance.

So considering above points, one possible solution can be:-

You can create a company.xsd file like:-

<?xml version="1.0" encoding="UTF-8"?>
<schema xmlns="http://www.w3.org/2001/XMLSchema"
    targetNamespace="http://www.example.org/NewXMLSchema"
    xmlns:tns="http://www.example.org/NewXMLSchema"
    elementFormDefault="unqualified">
    <element name="company">
        <complexType>
        <sequence>
            <element name="staff" type="tns:stafftype"/>
            </sequence>
        </complexType>

    </element>

    <complexType name="stafftype">
        <sequence>
        <element name="firstname" type="string" minOccurs="0" />
        <element name="lastname" type="string" minOccurs="0" />
        <element name="nickname" type="string" minOccurs="1" />
        <element name="salary" type="int" minOccurs="0" />
        </sequence>

    </complexType>

</schema>

then your java code would look like:-

import java.io.BufferedReader;
import java.io.ByteArrayInputStream;
import java.io.File;
import java.io.IOException;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

import javax.xml.transform.stream.StreamSource;
import javax.xml.validation.Schema;
import javax.xml.validation.SchemaFactory;
import javax.xml.validation.Validator;

import org.xml.sax.SAXException;

public class testXML {
    //  Lookup a factory for the W3C XML Schema language
    static SchemaFactory factory = SchemaFactory
            .newInstance("http://www.w3.org/2001/XMLSchema");

    //  Compile the schema. 
    static File schemaLocation = new File("company.xsd");
    static Schema schema = null;
    static {
        try {
            schema = factory.newSchema(schemaLocation);
        } catch (SAXException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
    }

    private final ExecutorService pool = Executors.newFixedThreadPool(20);;

    boolean validate(StringBuffer splitBuffer) {
        boolean isValid = false;
        Validator validator = schema.newValidator();
        try {
            validator.validate(new StreamSource(new ByteArrayInputStream(
                    splitBuffer.toString().getBytes())));
            isValid = true;
        } catch (SAXException ex) {
            System.out.println(ex.getMessage());
        } catch (IOException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }
        return isValid;

    }

    void split(BufferedReader br, String rootElementName,
            String splitElementName) {
        StringBuffer splitBuffer = null;
        String line = null;
        String startRootElement = "<" + rootElementName + ">";
        String endRootElement = "</" + rootElementName + ">";

        String startSplitElement = "<" + splitElementName + ">";
        String endSplitElement = "</" + splitElementName + ">";
        String xmlDeclaration = "<?xml version=\"1.0\"";
        boolean startFlag = false, endflag = false;
        try {
            while ((line = br.readLine()) != null) {
                if (line.contains(xmlDeclaration)
                        || line.contains(startRootElement)
                        || line.contains(endRootElement)) {
                    continue;
                }

                if (line.contains(startSplitElement)) {
                    startFlag = true;
                    endflag = false;
                    splitBuffer = new StringBuffer(startRootElement);
                    splitBuffer.append(line);

                } else if (line.contains(endSplitElement)) {
                    endflag = true;
                    startFlag = false;
                    splitBuffer.append(line);
                    splitBuffer.append(endRootElement);

                } else if (startFlag) {
                    splitBuffer.append(line);
                }

                if (endflag) {
                    //process splitBuffer
                    boolean result = validate(splitBuffer);
                    if (result) {
                        //send it to a thread for processing further
                        //it is async so that main thread can continue for next

                        pool.submit(new ProcessingHandler(splitBuffer));

                    }
                }

            }
        } catch (IOException e) {
            // TODO Auto-generated catch block
            e.printStackTrace();
        }

    }
}

class ProcessingHandler implements Runnable {
    String splitXML = null;

    ProcessingHandler(StringBuffer splitXMLBuffer) {
        this.splitXML = splitXMLBuffer.toString();
    }

    @Override
    public void run() {
        // do like writing to a file etc.

    }

}

Answer 6

Have a look at this. This is slightly reworked sample from xmlpull.org:

http://www.xmlpull.org/v1/download/unpacked/doc/quick_intro.html

The following should do all you need unless you have nested splitting tags like:

<?xml version="1.0" encoding="UTF-8" standalone="no" ?>
<company>
    <staff id="1">
        <firstname>yong</firstname>
        <lastname>mook kim</lastname>
        <nickname>mkyong</nickname>
        <salary>100000</salary>
        <other>
            <staff>
            ...
            </staff>
        </other>
    </staff>
</company>

To run it in pass-through mode simply pass null as splitting tag.

import java.io.File;
import java.io.FileNotFoundException;
import java.io.FileReader;
import java.io.IOException;

import org.apache.commons.io.FileUtils;
import org.xmlpull.v1.XmlPullParser;
import org.xmlpull.v1.XmlPullParserException;
import org.xmlpull.v1.XmlPullParserFactory;

public class XppSample {

private String rootTag;
private String splitTag;
private String requiredTag;
private int flushThreshold;
private String fileName;

private String rootTagEnd;

private boolean hasRequiredTag = false;
private int flushCount = 0;
private int fileNo = 0;
private String header;
private XmlPullParser xpp;
private StringBuilder nodeBuf = new StringBuilder();
private StringBuilder fileBuf = new StringBuilder();


public XppSample(String fileName, String rootTag, String splitTag, String requiredTag, int flushThreshold) throws XmlPullParserException, FileNotFoundException {

    this.rootTag = rootTag;
    rootTagEnd = "</" + rootTag + ">";
    this.splitTag = splitTag;
    this.requiredTag = requiredTag;
    this.flushThreshold = flushThreshold;
    this.fileName = fileName; 

    XmlPullParserFactory factory = XmlPullParserFactory.newInstance(System.getProperty(XmlPullParserFactory.PROPERTY_NAME), null);
    factory.setNamespaceAware(true);
    xpp = factory.newPullParser();
    xpp.setInput(new FileReader(fileName));
}


public void processDocument() throws XmlPullParserException, IOException {
    int eventType = xpp.getEventType();
    do {
        if(eventType == XmlPullParser.START_TAG) {
            processStartElement(xpp);
        } else if(eventType == XmlPullParser.END_TAG) {
            processEndElement(xpp);
        } else if(eventType == XmlPullParser.TEXT) {
            processText(xpp);
        }
        eventType = xpp.next();
    } while (eventType != XmlPullParser.END_DOCUMENT);

    saveFile();
}


public void processStartElement(XmlPullParser xpp) {

    int holderForStartAndLength[] = new int[2];
    String name = xpp.getName();
    char ch[] = xpp.getTextCharacters(holderForStartAndLength);
    int start = holderForStartAndLength[0];
    int length = holderForStartAndLength[1];

    if(name.equals(rootTag)) {
        int pos = start + length;
        header = new String(ch, 0, pos);
    } else {
        if(requiredTag==null || name.equals(requiredTag)) {
            hasRequiredTag = true;
        }
        nodeBuf.append(xpp.getText());
    }
}


public void flushBuffer() throws IOException {
    if(hasRequiredTag) {
        fileBuf.append(nodeBuf);
        if(((++flushCount)%flushThreshold)==0) {
            saveFile();
        }           
    }
    nodeBuf = new StringBuilder();
    hasRequiredTag = false;
}


public void saveFile() throws IOException {
    if(fileBuf.length()>0) {
        String splitFile = header + fileBuf.toString() + rootTagEnd;
        FileUtils.writeStringToFile(new File((fileNo++) + "_" + fileName), splitFile);
        fileBuf = new StringBuilder();
    }
}


public void processEndElement (XmlPullParser xpp) throws IOException {

    String name = xpp.getName();

    if(name.equals(rootTag)) {
        flushBuffer();
    } else {
        nodeBuf.append(xpp.getText());
        if(name.equals(splitTag)) {
            flushBuffer();
        }
    }
}


public void processText (XmlPullParser xpp) throws XmlPullParserException {

    int holderForStartAndLength[] = new int[2];
    char ch[] = xpp.getTextCharacters(holderForStartAndLength);
    int start = holderForStartAndLength[0];
    int length = holderForStartAndLength[1];
    String content = new String(ch, start, length);

    nodeBuf.append(content);
}


public static void main (String args[]) throws XmlPullParserException, IOException {

    //XppSample app = new XppSample("input.xml", "company", "staff", "nickname", 3);
    XppSample app = new XppSample("input.xml", "company", "staff", null, 3);
    app.processDocument();
}

}

Answer 7

Normally I would suggest using StAX, but it is unclear to me how 'stateful' your real XML is. If simple, then use SAX for ultimate performance, if not-so-simple, use StAX. So you need to

1. read bytes from disk
2. convert them to characters
3. parse the XML
4. determine whether to keep XML or throw away (skip out subtree)
5. write XML
6. convert characters to bytes
7. write to disk

Now, it might seem like steps 3-5 are the most resource-intensive, but I would rate them as

Most: 1 + 7
Middle: 2 + 6
Least: 3 + 4 + 5

As operations 1 and 7 are kind of seperate of the rest, you should do them in an async way, at least creating multiple small files is best done in n other threads, if you are familiar with multi-threading . For increased performance, you might also look into the new IO stuff in Java.

Now for steps 2 + 3 and 5 + 6 you can go a long way with FasterXML , it really does a lot of the stuff you are looking for, like triggering JVM hot-spot attention in the right places; might even support async reading/writing looking through the code quickly.

So then we are left with step 5, and depending on your logic, you should either

a. make an object binding, then decide how what to do
b. write XML anyways, hoping for the best, and then throw it away if no 'staff' element is present.

Whatever you do, object reuse is sensible. Note that both alternatives (obisously) requires the same amount of parsing (skip out of subtree ASAP), and for alternative b, that a little extra XML is actually not so bad performancewise, ideally make sure your char buffers are > one unit.

Alternative b is the most easy to implement, simply copy the 'xml event' from your reader to writer, example for StAX:

private static void copyEvent(int event, XMLStreamReader  reader, XMLStreamWriter writer) throws XMLStreamException {
    if (event == XMLStreamConstants.START_ELEMENT) {
        String localName = reader.getLocalName();
        String namespace = reader.getNamespaceURI();
        // TODO check this stuff again before setting in production
        if (namespace != null) {
            if (writer.getPrefix(namespace) != null) {
                writer.writeStartElement(namespace, localName);
            } else {
                writer.writeStartElement(reader.getPrefix(), localName, namespace);
            }
        } else {
            writer.writeStartElement(localName);
        }
        // first: namespace definition attributes
        if(reader.getNamespaceCount() > 0) {
            int namespaces = reader.getNamespaceCount();

            for(int i = 0; i < namespaces; i++) {
                String namespaceURI = reader.getNamespaceURI(i);

                if(writer.getPrefix(namespaceURI) == null) {
                    String namespacePrefix = reader.getNamespacePrefix(i);

                    if(namespacePrefix == null) {
                        writer.writeDefaultNamespace(namespaceURI);
                    } else {
                        writer.writeNamespace(namespacePrefix, namespaceURI);
                    }
                }
            }
        }
        int attributes = reader.getAttributeCount();

        // the write the rest of the attributes
        for (int i = 0; i < attributes; i++) {
            String attributeNamespace = reader.getAttributeNamespace(i);
            if (attributeNamespace != null && attributeNamespace.length() != 0) {
                writer.writeAttribute(attributeNamespace, reader.getAttributeLocalName(i), reader.getAttributeValue(i));
            } else {
                writer.writeAttribute(reader.getAttributeLocalName(i), reader.getAttributeValue(i));
            }
        }
    } else if (event == XMLStreamConstants.END_ELEMENT) {
        writer.writeEndElement();
    } else if (event == XMLStreamConstants.CDATA) {
        String array = reader.getText();
        writer.writeCData(array);
    } else if (event == XMLStreamConstants.COMMENT) {
        String array = reader.getText();
        writer.writeComment(array);
    } else if (event == XMLStreamConstants.CHARACTERS) {
        String array = reader.getText();
        if (array.length() > 0 && !reader.isWhiteSpace()) {
            writer.writeCharacters(array);
        }
    } else if (event == XMLStreamConstants.START_DOCUMENT) {
        writer.writeStartDocument();
    } else if (event == XMLStreamConstants.END_DOCUMENT) {
        writer.writeEndDocument();
    }
}

And for a subtree,

private static void copySubTree(XMLStreamReader reader, XMLStreamWriter writer) throws XMLStreamException {
    reader.require(XMLStreamConstants.START_ELEMENT, null, null);

    copyEvent(XMLStreamConstants.START_ELEMENT, reader, writer);

    int level = 1;
    do {
        int event = reader.next();
        if(event == XMLStreamConstants.START_ELEMENT) {
            level++;
        } else if(event == XMLStreamConstants.END_ELEMENT) {
            level--;
        }

        copyEvent(event, reader, writer);
    } while(level > 0);

}

From which you probably can deduct how to skip out to a certain level. In general, for stateful StaX parsing, use the pattern

private static void parseSubTree(XMLStreamReader reader) throws XMLStreamException {

    int level = 1;
    do {
        int event = reader.next();
        if(event == XMLStreamConstants.START_ELEMENT) {
            level++;
            // do stateful stuff here

            // for child logic:
            if(reader.getLocalName().equals("Whatever")) {
                parseSubTreeForWhatever(reader);
                level --; // read from level 1 to 0 in submethod.
            }

            // alternatively, faster
            if(level == 4) {
                parseSubTreeForWhateverAtRelativeLevel4(reader);
                level --; // read from level 1 to 0 in submethod.
            }


        } else if(event == XMLStreamConstants.END_ELEMENT) {
            level--;
            // do stateful stuff here, too
        }

    } while(level > 0);

}

where you in the start of the document read till the first start element and break (add the writer+copy for your use of course, as above).

Note that if you do an object binding, these methods should be placed in that object, and equally for the serialization methods.

I am pretty sure you will get 10s of MB/s on a modern system, and that should be sufficient. An issue to be investigate further, is approaches to use multiple cores for the actualy input, if you know for a fact the encoding subset, like non-crazy UTF-8, or ISO-8859, then random access might be possible -> send to different cores.

Have fun, and tell use how it went ;)

: Almost forgot, if you for some reason are the one who is creating the file in the first place, or you will be reading them after splitting, you will se HUGE performance gains using XML binarization; there exist XML Schema generators which again can go into code generators. (And some XSLT transform libs use code generation too.) And run with the -server option for JVM.

Answer 8

My suggestion is that SAX, STAX, or DOM are not the ideal xml parser for your problem, the perfect solutions is called vtd-xml , there is an article on this subject explaining why DOM sax and STAX all done something very wrong... the code below is the shortest you have to write, yet performs 10x faster than DOM or SAX. http://www.javaworld.com/javaworld/jw-07-2006/jw-0724-vtdxml.html

Here is a latest paper entitled Processing XML with Java – A Performance Benchmark : http://recipp.ipp.pt/bitstream/10400.22/1847/1/ART_BrunoOliveira_2013.pdf

import com.ximpleware.*;
import java.io.*;
public class gandalf {
    public  static void main(String a[]) throws VTDException, Exception{
        VTDGen vg = new VTDGen();
        if (vg.parseFile("c:\\xml\\gandalf.txt", false)){
            VTDNav vn=vg.getNav();
            AutoPilot ap = new AutoPilot(vn);
            ap.selectXPath("/company/staff[nickname]");
            int i=-1;
            int count=0;
            while((i=ap.evalXPath())!=-1){
                vn.dumpFragment("c:\\xml\\staff"+count+".xml");
                count++;
            }
        }
    }

}

Answer 9

How to make i faster:

1. Use asynchronous writes, possibly in parallel, might boost your perf if you have RAID-X something disks
2. Write to an SSD instead of HDD

Answer 10

Here is DOM based solution. I have tested this with the xml you provided. This needs to be checked against the actual xml files that you have.

Since this is based on DOM parser, please remember that this will require a lot of memory depending upon your xml file size . But its much faster as it's DOM based.

Algorithm :

1. Parse the document
2. Extract the root element name
3. Get list he nodes based on the split criteria (using XPath)
4. For each node, create an empty document with root element name as extracted in step #2
5. Insert the node in this new document
6. Check if nodes are to be filtered or not.
7. If nodes are to be filtered, then check if a specified element is present in the newly created doc.
8. If node is not present, don't write to the file.
9. If the nodes are NOT to be filtered at all, don't check for condition in #7, and write the document to the file.

This can be run from command prompt as follows

java    XMLSplitter xmlFileLocation  splitElement filter filterElement

For the xml you mentioned it will be

java    XMLSplitter input.xml  staff  true nickname

In case you don't want to filter

java    XMLSplitter input.xml  staff 

Here is the complete java code:

package com.xml.xpath;

import java.io.File;
import java.io.FileWriter;
import java.io.IOException;
import java.io.StringReader;
import java.io.StringWriter;

import javax.xml.parsers.DocumentBuilder;
import javax.xml.parsers.DocumentBuilderFactory;
import javax.xml.parsers.ParserConfigurationException;
import javax.xml.transform.OutputKeys;
import javax.xml.transform.Transformer;
import javax.xml.transform.TransformerConfigurationException;
import javax.xml.transform.TransformerException;
import javax.xml.transform.TransformerFactory;
import javax.xml.transform.dom.DOMSource;
import javax.xml.transform.stream.StreamResult;
import javax.xml.xpath.XPath;
import javax.xml.xpath.XPathConstants;
import javax.xml.xpath.XPathExpression;
import javax.xml.xpath.XPathExpressionException;
import javax.xml.xpath.XPathFactory;

import org.w3c.dom.DOMException;
import org.w3c.dom.DOMImplementation;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;
import org.xml.sax.InputSource;
import org.xml.sax.SAXException;

public class XMLSplitter {

    DocumentBuilder builder = null;
    XPath xpath = null; 
    Transformer transformer = null;
    String filterElement;
    String splitElement;
    String xmlFileLocation;
    boolean filter = true;


    public static void main(String[] arg) throws Exception{

        XMLSplitter xMLSplitter = null;
        if(arg.length < 4){

            if(arg.length < 2){
                System.out.println("Insufficient arguments !!!");
                System.out.println("Usage: XMLSplitter xmlFileLocation  splitElement filter filterElement ");
                return;
            }else{
                System.out.println("Filter is off...");
                xMLSplitter = new XMLSplitter();
                xMLSplitter.init(arg[0],arg[1],false,null);
            }

        }else{
            xMLSplitter = new XMLSplitter();
            xMLSplitter.init(arg[0],arg[1],Boolean.parseBoolean(arg[2]),arg[3]);
        }



        xMLSplitter.start();    

    }

    public void init(String xmlFileLocation, String splitElement, boolean filter, String filterElement ) 
                throws ParserConfigurationException, TransformerConfigurationException{

        //Initialize the Document builder
        System.out.println("Initializing..");
        DocumentBuilderFactory domFactory = DocumentBuilderFactory.newInstance();
        domFactory.setNamespaceAware(true);   
        builder = domFactory.newDocumentBuilder();

        //Initialize the transformer
        TransformerFactory transformerFactory = TransformerFactory.newInstance();
        transformer = transformerFactory.newTransformer();
        transformer.setOutputProperty(OutputKeys.METHOD, "xml");
        transformer.setOutputProperty(OutputKeys.ENCODING,"UTF-8");
        transformer.setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "4");
        transformer.setOutputProperty(OutputKeys.INDENT, "yes");

        //Initialize the xpath
        XPathFactory factory = XPathFactory.newInstance();
        xpath = factory.newXPath();

        this.filterElement = filterElement;
        this.splitElement = splitElement;
        this.xmlFileLocation = xmlFileLocation;
        this.filter = filter;


    }   


    public void start() throws Exception{

            //Parser the file 
            System.out.println("Parsing file.");
            Document doc = builder. parse(xmlFileLocation);

            //Get the root node name
            System.out.println("Getting root element.");
            XPathExpression rootElementexpr = xpath.compile("/");
            Object rootExprResult = rootElementexpr.evaluate(doc, XPathConstants.NODESET);
            NodeList rootNode = (NodeList) rootExprResult;          
            String rootNodeName = rootNode.item(0).getFirstChild().getNodeName();

            //Get the list of split elements
            XPathExpression expr = xpath.compile("//"+splitElement);
            Object result = expr.evaluate(doc, XPathConstants.NODESET);
            NodeList nodes = (NodeList) result;
            System.out.println("Total number of split nodes "+nodes.getLength());
            for (int i = 0; i < nodes.getLength(); i++) {
                //Wrap each node inside root of the parent xml doc
                Node sigleNode = wrappInRootElement(rootNodeName,nodes.item(i));
                //Get the XML string of the fragment
                String xmlFragment = serializeDocument(sigleNode);
                //System.out.println(xmlFragment);
                //Write the xml fragment in file.
                storeInFile(xmlFragment,i);         
            }

    }

    private  Node wrappInRootElement(String rootNodeName, Node fragmentDoc) 
                throws XPathExpressionException, ParserConfigurationException, DOMException, 
                        SAXException, IOException, TransformerException{

        //Create empty doc with just root node
        DOMImplementation domImplementation = builder.getDOMImplementation();
        Document doc = domImplementation.createDocument(null,null,null);
        Element theDoc = doc.createElement(rootNodeName);
        doc.appendChild(theDoc);

        //Insert the fragment inside the root node 
        InputSource inStream = new InputSource();     
        String xmlString = serializeDocument(fragmentDoc);
        inStream.setCharacterStream(new StringReader(xmlString));       
        Document fr = builder.parse(inStream);
        theDoc.appendChild(doc.importNode(fr.getFirstChild(),true));
        return doc;
    }

    private String serializeDocument(Node doc) throws TransformerException, XPathExpressionException{

        if(!serializeThisNode(doc)){
            return null;
        }

        DOMSource domSource = new DOMSource(doc);                
        StringWriter stringWriter = new StringWriter();
        StreamResult streamResult = new StreamResult(stringWriter);
        transformer.transform(domSource, streamResult);
        String xml = stringWriter.toString();
        return xml;

    }

    //Check whether node is to be stored in file or rejected based on input
    private boolean serializeThisNode(Node doc) throws XPathExpressionException{

         if(!filter){
             return true;
         }

         XPathExpression filterElementexpr = xpath.compile("//"+filterElement);
         Object result = filterElementexpr.evaluate(doc, XPathConstants.NODESET);
         NodeList nodes = (NodeList) result;

         if(nodes.item(0) != null){
             return true;
         }else{
             return false;
         }       
    }

    private void storeInFile(String content, int fileIndex) throws IOException{

        if(content == null || content.length() == 0){
            return;
        }

        String fileName = splitElement+fileIndex+".xml";

        File file = new File(fileName);
        if(file.exists()){
            System.out.println(" The file "+fileName+" already exists !! cannot create the file with the same name ");
            return;
        }
        FileWriter fileWriter = new FileWriter(file);
        fileWriter.write(content);
        fileWriter.close();
        System.out.println("Generated file "+fileName);


    }

}

Let me know if this works for you or any other help regarding this code.