Regular expressions in XML schemas

adrianbarry · March 26th, 2004, 11:55 AM

Hello,

I am trying to validate an XML document using an XML schema. I have a piece of data that I am trying to validate by using a regular expression (patterned type). My data can be an empty string, but if it is present it must conform to the pattern a999999, the letter a followed by 6 digits. Can anybody tell me how to achieve this? I tried the code below, but it does not allow for an empty string.

<element name="agentid" nillable="true">
    <simpleType>
        <restriction base="string">
            <pattern value="a\d{6}"/>
        </restriction>
    </simpleType>
</element>

Many thanks

Adrian

joefawcett · March 26th, 2004, 01:26 PM

I think you can use the union operator:
value="(a\d{6})|()"
no guarantees as I'm unable to test this at the moment.

--

Joe

adrianbarry · March 29th, 2004, 05:06 AM

Hi Joe,

Thanks for your suggestion. Unfortunately it didn't work. My schema validator (xmlspy version 5 rel. 4) does not think that the pattern type is valid. The search continues...

Thanks

Adrian

joefawcett · March 29th, 2004, 05:47 AM

Unfortunately Schemas don't suppot all the atndard patterns. I will have a look at my Schema bible (Definitive XML Schemas published by Prentice Hall).

--

Joe

joefawcett · March 29th, 2004, 05:55 AM

Are you specifying the element to be empty in the instance by using xsi:nil="true"? It is not enough to be just empty.

--

Joe

adrianbarry · March 29th, 2004, 06:48 AM

Tried that. Still don't get the choice of empty or matching a999999.

Adrian

joefawcett · March 29th, 2004, 06:57 AM

What parser and validator are you using?

--

Joe

adrianbarry · March 29th, 2004, 09:49 AM

I'm using XMLSPY version 5 release 4

Adrian

joefawcett · March 31st, 2004, 01:28 PM

I don't have that but here is a test bed, six filkes all told, save in same directory and run ValidateXml.js:

Code:


<xsd:schema 
     xmlns:xsd="http://www.w3.org/2001/XMLSchema"
     xmlns:tns="http://tempuri.org/test.xsd"
     targetNamespace="http://tempuri.org/test.xsd"
      elementFormDefault="qualified">    
    <xsd:element name="agentid" nillable="true">
      <xsd:simpleType>
        <xsd:restriction base="xsd:string">
          <xsd:pattern value="a\d{6}"/>
        </xsd:restriction>
     </xsd:simpleType>
    </xsd:element>
  </xsd:schema>



  <agentid xmlns="http://tempuri.org/test.xsd"/>



  <agentid xmlns="http://tempuri.org/test.xsd">a123456</agentid>



  <agentid xmlns="http://tempuri.org/test.xsd">a123456</agentid>



  <agentid xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://tempuri.org/test.xsd" xsi:nil="true"/>  


  var sSchemaPath = "nillable.xsd";
  var sTns = "http://tempuri.org/test.xsd";
  var arrTestDocs = ["nillable1.xml", "nillable2.xml", "nillable3.xml", "nillable4.xml"];

  function getAsyncDom()
  {
    var oDom =  new ActiveXObject("Msxml2.DomDocument.4.0");
    oDom.async = false;
    return oDom
  }

  function getSchemaCache()
  {
    return new ActiveXObject("Msxml2.XMLSchemaCache.4.0");
  }

  function doValidation()
  {
    var oSC = getSchemaCache();
    oSC.add(sTns, sSchemaPath);
    var oDom = getAsyncDom();
    oDom.validateOnParse = false;
    for (var i = 0; i < arrTestDocs.length; i++)
    {
      oDom.load(arrTestDocs[i]);
      oDom.schemas = oSC;
      var oError = oDom.validate();
      if (oError.errorCode == 0)
      {
        WScript.echo(arrTestDocs[i] + " passes"); 
      }
      else
      {
        WScript.echo(arrTestDocs[i] + " fails:\n" + oError.reason + "\n" + oError.srcText);    
      }  
    }
    oDom = null;
    oSC = null;
  }

  doValidation();

--

Joe

mhkay · April 6th, 2004, 01:55 PM

Can't you just use "(a\d{6})?" (including the question mark)?

Michael Kay
http://saxon.sf.net/