Question

I have an XML file, I need to extract values from it, and put them in another XML file.

Questions:

1. Another person is creating the “schema” for the resulting XML file. Is there something that person can give me that will automate the inserting of the values? Do I even need to extract anything from the XML, or can something like a XSLT just do all the transformation?

2. Is there a problem with this XML structure below? I tried using xsd2code to generate objects but nothing will load when I use the LoadFromFileMethod – I read an article that wasn’t very specific but said “nested parents” cause problems for XSD.exe and xsd2code.

<Section>
<Form id="1"...>
<Control id="12523"..> <--Some have this some don't
  <Property name="Color">Red</Property>
  <Property name="Size">Large</Property>
</Control>
</Form>
<Form id="2"...>
  <Property name="Color">Blue</Property>
  <Property name="Size">Large</Property>
</Form>
<Form id="3"...>
  <Property name="Color">Red</Property>
  <Property name="Size">Small</Property>
</Form>
</Section>

Thank you for any guidance!

Answer 1

XSLT is the tool for XML transformation.

As far as your XML goes, in a lot of applications you should replace this:

<Property name="Color">Red</Property>

with:

<Color>Red</Color>

Some reasons:

1. If you want to write a schema that restricts an element’s content in some way (e.g. to one of a list of values), the element must be identifiable by its name; you can’t write one schema for a Property element with a name attribute equals “Color” and another schema for a Property element whose name attribute equals “Size”.

2. It’s easier to write XPath predicates if your element names are meaningful. For instance, Form[Color = 'Red'] is a lot easier to write (and read) than Form[Property[@name='Color' and .='Red']]

3. The above is also true if you’re writing Linq queries against the XML, in pretty much the same manner. Compare Element.Descendants("Color") with Element.Descendents("Property").Where(x => x.Attributes["name"] == "Color").

There are applications where it’s appropriate to have generically-named elements, too; the above argument’s not definitive. But if you’re going to do that, you should have good reasons.