«

I'm trying to figure out exactly what's going wrong with an xpath
keyref validation scenario.

In brief, I have something like the following in my xsd:
<xs:schema>
<xs:complexType name="topLevelContainer">

<!-- define an element -->
<xs:complexType name="element">
...
<xs:attribute name="id" type="xs:int" use="required"/>
</xs:complexType>

<!-- instance one, simple list -->
<xs:element name="instanceOne" minOccurs="0">
<xs:sequecne>
<xs:element name="elementDescription" type="element"
minOccurs="0" maxOccurs="unbounded"/>
</xs:sequence>
</xs:complexType>

<!-- instance two, an inner, heterogenous list -->
<xs:complexType name="instanceTwo" type="anotherType">
<xs:element name="elementList">
<xs:complexType>
<xs:choice maxOccurs="unbounded">
<xs:complexType name="fullElement"
type="elementType"/>
<xs:complexType name="elementRef" type="xs:int"/>
</xs:choice>
</xs:complexType>
</xs:sequence>

<!-- constraints -->
<!-- key points to the id attribute of instanceOne type -->
<xs:key name="elementKey">
<xs:selector xpath="instanceOne/elementDescription"/>
<xs:field xpath="@id">
</xs:key>

<!-- keyref points to more-deeply nested instanceTwo type -->
<xs:keyref name="elementRef" refer="elementKey">
<xs:selector xpath="instanceTwo/elementList/elementRef"/>
<xs:field xpath="."/>
</xs:keyref>

</xs:complexType> <!-- end top-level container -->
</xs:schema>

In short, what I am trying to do is ensure that the references from
within the instanceTwo type point to elements within the instanceOne
type. But what I'm not able to figure out is how to enforce that the
instanceTwo type contain references to fully described elements in
instanceOne only. For some reason, a file will validate when the
references out of instanceTwo refer to elementDescriptions from other
instanceTwos in the file.

I'm suspicious that the xpath comparison will never return false for
some reason related to datatype. (My xml editor seems to suggest that
the xpath expressions are returning the appropriate values.)

???

Hi,

Your schema has quite a large number of structural problems - keys
within complex types instead of elements, complex types with type
attributes, complex types directly within choice groups, etc.

I'm assuming this is because you cut down the example to be concise, but
it makes it awfully hard to debug the problem when it's not a valid
schema.

Your general approach looks right to me. It could be a namespace issue
- if you're using a target namespace, you must prefix all your element
names in your XPaths. It could also be a problem with your XML Editor,
since some of them are buggy when it comes to identity constraints.

Otherwise, you could post the real schema and we could take a look at
it.

Hope that helps,
Priscilla

----------------------------------
Priscilla Walmsley
Author, Definitive XML Schema
Definitive XQuery
http://www.datypic.com
----------------------------------

*** Sent via Developersdex http://www.developersdex.com ***

Hi,

The xsd is available at
http://psidev.sourceforge.net/mi/rel25/src/MIF25.xsd . XML Spy
documentation is available at
http://psidev.sourceforge.net/mi/rel25/doc/ . As you mentioned it is
quite complicated and likely to have some structural problems, and I've
made some minor revisions to the original for purposes of the example
as I've analyzed the problem.

Thanks...

Hi,

The most obvious thing I notice is that you are using a target
namespace, but you are not prefixing the element names in your paths.
XPath expressions are not affected by default namespace declarations, so
you will need to declare a prefix for the default namespace, e.g.

xmlnsyz="net:sfsidev:mi"

Then prefix all the element names in your XPaths, e.g.

<xs:selector xpath="xyz:experimentList/xyz:experimentDescription" />

Hope that helps,
Priscilla
----------------------------------
Priscilla Walmsley
Author, Definitive XML Schema
Definitive XQuery
http://www.datypic.com
----------------------------------

*** Sent via Developersdex http://www.developersdex.com ***

Hi Priscilla,

Perhaps you could suggest a good xml style guide or a tool better than
those that I am using? Both XmlSpy and a java validator I wrote myself
(sax parser) fail to choke on the problem use case that I have on hand.

Thanks again...
Jason