现在已经是2012年了，Java可以比以前的XML做更多的事情，我想在我公认的答案之外添加一个替代方案。这在Java 6之外没有依赖关系。

import org.w3c.dom.Node;
import org.w3c.dom.bootstrap.DOMImplementationRegistry;
import org.w3c.dom.ls.DOMImplementationLS;
import org.w3c.dom.ls.LSSerializer;
import org.xml.sax.InputSource;

import javax.xml.parsers.DocumentBuilderFactory;
import java.io.StringReader;

/**
 * Pretty-prints xml, supplied as a string.
 * <p/>
 * eg.
 * <code>
 * String formattedXml = new XmlFormatter().format("<tag><nested>hello</nested></tag>");
 * </code>
 */
public class XmlFormatter {

    public String format(String xml) {

        try {
            final InputSource src = new InputSource(new StringReader(xml));
            final Node document = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(src).getDocumentElement();
            final Boolean keepDeclaration = Boolean.valueOf(xml.startsWith("<?xml"));

        //May need this: System.setProperty(DOMImplementationRegistry.PROPERTY,"com.sun.org.apache.xerces.internal.dom.DOMImplementationSourceImpl");


            final DOMImplementationRegistry registry = DOMImplementationRegistry.newInstance();
            final DOMImplementationLS impl = (DOMImplementationLS) registry.getDOMImplementation("LS");
            final LSSerializer writer = impl.createLSSerializer();

            writer.getDomConfig().setParameter("format-pretty-print", Boolean.TRUE); // Set this to true if the output needs to be beautified.
            writer.getDomConfig().setParameter("xml-declaration", keepDeclaration); // Set this to true if the declaration is needed to be outputted.

            return writer.writeToString(document);
        } catch (Exception e) {
            throw new RuntimeException(e);
        }
    }

    public static void main(String[] args) {
        String unformattedXml =
                "<?xml version=\"1.0\" encoding=\"UTF-8\"?><QueryMessage\n" +
                        "        xmlns=\"http://www.SDMX.org/resources/SDMXML/schemas/v2_0/message\"\n" +
                        "        xmlns:query=\"http://www.SDMX.org/resources/SDMXML/schemas/v2_0/query\">\n" +
                        "    <Query>\n" +
                        "        <query:CategorySchemeWhere>\n" +
                        "   \t\t\t\t\t         <query:AgencyID>ECB\n\n\n\n</query:AgencyID>\n" +
                        "        </query:CategorySchemeWhere>\n" +
                        "    </Query>\n\n\n\n\n" +
                        "</QueryMessage>";

        System.out.println(new XmlFormatter().format(unformattedXml));
    }
}

有一个非常好的命令行XML实用程序叫做xmlstarlet(http://xmlstar.sourceforge.net/)，它可以做很多事情，很多人都在使用它。

您可以使用Runtime以编程方式执行此程序。然后读入格式化的输出文件。它具有比几行Java代码所能提供的更多选项和更好的错误报告。

下载xmlstarlet: http://sourceforge.net/project/showfiles.php?group_id=66612&package_id=64589

Transformer transformer = TransformerFactory.newInstance().newTransformer();
transformer.setOutputProperty(OutputKeys.INDENT, "yes");
transformer.setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "2");
// initialize StreamResult with File object to save to file
StreamResult result = new StreamResult(new StringWriter());
DOMSource source = new DOMSource(doc);
transformer.transform(source, result);
String xmlString = result.getWriter().toString();
System.out.println(xmlString);

注意:根据Java版本的不同，结果可能有所不同。搜索特定于您的平台的解决方案。

Since you are starting with a String, you can convert to a DOM object (e.g. Node) before you use the Transformer. However, if you know your XML string is valid, and you don't want to incur the memory overhead of parsing a string into a DOM, then running a transform over the DOM to get a string back - you could just do some old fashioned character by character parsing. Insert a newline and spaces after every </...> characters, keep and indent counter (to determine the number of spaces) that you increment for every <...> and decrement for every </...> you see.

免责声明-我对下面的函数做了剪切/粘贴/文本编辑，所以它们可能不能按原样编译。

public static final Element createDOM(String strXML) 
    throws ParserConfigurationException, SAXException, IOException {

    DocumentBuilderFactory dbf = DocumentBuilderFactory.newInstance();
    dbf.setValidating(true);
    DocumentBuilder db = dbf.newDocumentBuilder();
    InputSource sourceXML = new InputSource(new StringReader(strXML));
    Document xmlDoc = db.parse(sourceXML);
    Element e = xmlDoc.getDocumentElement();
    e.normalize();
    return e;
}

public static final void prettyPrint(Node xml, OutputStream out)
    throws TransformerConfigurationException, TransformerFactoryConfigurationError, TransformerException {
    Transformer tf = TransformerFactory.newInstance().newTransformer();
    tf.setOutputProperty(OutputKeys.OMIT_XML_DECLARATION, "yes");
    tf.setOutputProperty(OutputKeys.ENCODING, "UTF-8");
    tf.setOutputProperty(OutputKeys.INDENT, "yes");
    tf.transform(new DOMSource(xml), new StreamResult(out));
}

I have found that in Java 1.6.0_32 the normal method to pretty print an XML string (using a Transformer with a null or identity xslt) does not behave as I would like if tags are merely separated by whitespace, as opposed to having no separating text. I tried using <xsl:strip-space elements="*"/> in my template to no avail. The simplest solution I found was to strip the space the way I wanted using a SAXSource and XML filter. Since my solution was for logging I also extended this to work with incomplete XML fragments. Note the normal method seems to work fine if you use a DOMSource but I did not want to use this because of the incompleteness and memory overhead.

public static class WhitespaceIgnoreFilter extends XMLFilterImpl
{

    @Override
    public void ignorableWhitespace(char[] arg0,
                                    int arg1,
                                    int arg2) throws SAXException
    {
        //Ignore it then...
    }

    @Override
    public void characters( char[] ch,
                            int start,
                            int length) throws SAXException
    {
        if (!new String(ch, start, length).trim().equals("")) 
               super.characters(ch, start, length); 
    }
}

public static String prettyXML(String logMsg, boolean allowBadlyFormedFragments) throws SAXException, IOException, TransformerException
    {
        TransformerFactory transFactory = TransformerFactory.newInstance();
        transFactory.setAttribute("indent-number", new Integer(2));
        Transformer transformer = transFactory.newTransformer();
        transformer.setOutputProperty(OutputKeys.INDENT, "yes");
        transformer.setOutputProperty("{http://xml.apache.org/xslt}indent-amount", "4");
        StringWriter out = new StringWriter();
        XMLReader masterParser = SAXHelper.getSAXParser(true);
        XMLFilter parser = new WhitespaceIgnoreFilter();
        parser.setParent(masterParser);

        if(allowBadlyFormedFragments)
        {
            transformer.setErrorListener(new ErrorListener()
            {
                @Override
                public void warning(TransformerException exception) throws TransformerException
                {
                }

                @Override
                public void fatalError(TransformerException exception) throws TransformerException
                {
                }

                @Override
                public void error(TransformerException exception) throws TransformerException
                {
                }
            });
        }

        try
        {
            transformer.transform(new SAXSource(parser, new InputSource(new StringReader(logMsg))), new StreamResult(out));
        }
        catch (TransformerException e)
        {
            if(e.getCause() != null && e.getCause() instanceof SAXParseException)
            {
                if(!allowBadlyFormedFragments || !"XML document structures must start and end within the same entity.".equals(e.getCause().getMessage()))
                {
                    throw e;
                }
            }
            else
            {
                throw e;
            }
        }
        out.flush();
        return out.toString();
    }

我用Scala看到了一个答案，所以这里有另一个用Groovy的答案，以防有人觉得有趣。默认缩进为2步，XmlNodePrinter构造函数也可以传递另一个值。

def xml = "<tag><nested>hello</nested></tag>"
def stringWriter = new StringWriter()
def node = new XmlParser().parseText(xml);
new XmlNodePrinter(new PrintWriter(stringWriter)).print(node)
println stringWriter.toString()

如果groovy jar在类路径中，则使用Java

  String xml = "<tag><nested>hello</nested></tag>";
  StringWriter stringWriter = new StringWriter();
  Node node = new XmlParser().parseText(xml);
  new XmlNodePrinter(new PrintWriter(stringWriter)).print(node);
  System.out.println(stringWriter.toString());