how can I convert xsd: pattern in java regex

Question

As I know, and I used very little java regex, there is a method (or tool) to convert a control xsd:pattern in java regex?

My xsd: pattern is as follows:

<xsd:simpleType name="myCodex">
<xsd:restriction base="xsd:string">
 <xsd:pattern value="[A-Za-z]{6}[0-9]{2}[A-Za-z]{1}[0-9]{2}[A-Za-z]{1}[0-9A-Za-z]{3}[A-Za-z]{1}" />
 <xsd:pattern value="[A-Za-z]{6}[0-9LMNPQRSTUV]{2}[A-Za-z]{1}[0-9LMNPQRSTUV]{2}[A-Za-z]{1}[0-9LMNPQRSTUV]{3}[A-Za-z]{1}" />
 <xsd:pattern value="[0-9]{11,11}" />
</xsd:restriction>
</xsd:simpleType>

Answer 1

You can load the XSD into Java and extract the expressions. Then you can use them in .matches() methods or create Pattern objects if you are going to reuse them a lot.

First you need to load the XML into a Java program (I called it CodexSchema.xsd):

DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
DocumentBuilder builder = factory.newDocumentBuilder();
Document source = builder.parse(new File("CodexSchema.xsd"));

Then you can use XPath to find the patterns you want to extract (you might want to create a method that takes the name of the simple type, if you have many to process). I used a more complicated XPath expression to avoid registering the namespaces:

XPathFactory xPathfactory = XPathFactory.newInstance();
String typeName = "myCodex";
String xPathRoot = "//*[local-name()='simpleType'][@name='"+typeName+"']/*[local-name()='restriction']/*[local-name()='pattern']";
XPath patternsXPath = xPathfactory.newXPath(); // this represents the NodeList of <xs:pattern> elements

Running that expression you get org.xml.dom.NodeList containing the <xs:pattern> elements.

NodeList patternNodes = (NodeList)patternsXPath.evaluate(xPathRoot, source, XPathConstants.NODESET);

Now you can loop through them and extract the contents of their value attribute. You might want to write a method for that:

public List<Pattern> getPatterns(NodeList patternNodes) {
    List<Pattern> expressions = new ArrayList<>();
    for(int i = 0; i < patternNodes.getLength(); i++) {
        Element patternNode = (Element)patternNodes.item(i);
        String regex = patternNode.getAttribute("value");
        expressions.add(Pattern.compile(regex));
    }
    return expressions;
}

You don't really need to put them into Pattern. You could simply use String.

You can now read all your patterns in Java using:

for(Pattern p : getPatterns(patternNodes)) {
    System.out.println(p);
}

Here are some tests with the third pattern:

Pattern pattern3 = getPatterns(patternNodes).get(2);

Matcher matcher = pattern3.matcher("47385628403");
System.out.println("test1: " + matcher.find());  // prints `test1: true`

System.out.println("test2: " + "47385628403".matches(pattern3.toString()));  // prints `test2: true`