Re: Sanitizing data

David,

Sorry about not providing a sample, the data was sensitive; I 've tried to sum up the data here:

<page>
  <page_type>
    ...
    <page_title>
      <code>10</code>
      <value>Page title</value>
    </page_title>
    <page_kind>
      <code>11</code>
      <value>Z1</value>
    </page_kind>
    ...
  </page_type>
  <page_data>
    ...
    <page_prev>
      <page_before>
        <!-- page_ref name and count differs -->
        <page_ref>
          <code>62</code>
          <value1>ABC</value1>
          <value2>123</value2>
          <value3>XYZ</value3>
          ...
          <!-- A sub group of nodes can appear under here -->
          <valuex>
            ...
            <somevalues>1,2,3</somevalues>
            <someotherdata>a,b,c</someotherdata>
            ...
          </valuex>
        </page_ref>
      </page_before>
    </page_prev>
    ...
  </page_data>
</page>

I believe that effective use of the 'otherwise' processing instruction may save me here. My problem is that the value of 'page_ref/somevalues' could mean that I need to process 'page_title/value' differently. This is part of my problem, the source daata contains a mix of processing insructions alongside the data; which in turn affects part of the data which is in a completely path.

I think that by counting expected numbers of nodes and using 'otherwise' I should be able to trap all conditions that should not occur.

Thanks,
JV