CODEWITHSUNDEEP
xmlxsltxsd
ShadowFlame
ShadowFlame

Reputation: 3278

xslt being aware of xs:import in xsd

I have several xsd's that I do not own, that need to be converted to an in-house xml standard to work within our application.

There is an xslt script that does this for us through an ant task, but now we've come to a roadblock:
when the xsd contains an xs:import statement, the xslt will not be able to find the fields defined in the referenced xsd.

example xsd's:
main.xsd

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema>
    <xs:import namespace="urn:company:ns.common" schemaLocation="common.xsd"/> 
    <xs:complexType name="User">
        <xs:element ref="common:Address">
    </xs:complexType>
</xs:schema>

common.xsd

<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns="urn:company:ns.common">
   <xs:complexType name="Address">
   <xs:sequence>
      <xs:element name="Street" type="xs:string"/>
      <xs:element name="HouseNumber" type="xs:number"/>
  </xs:sequence>
 </xs:complexType>
</xs:schema>

while I can't show the current xslt since it would be akin to divulging some company information, the expected output should be something like this:

<User>
  <Address>
    <Street></Street>
    <HouseNumber></HouseNumber>
  </Address>
</User>

at this time, the xslt ignores what it can't find when it is inside the imported xsd.

wrong result:

<User></User>

Upvotes: 0

Views: 164

Answers (1)

kjhughes
kjhughes

Reputation: 111581

Without making one or more simplifying assumptions, you have a complex subproject on your hands, not a SO Q/A. I'll suggest a key simplifying assumption:

  • The separate XSDs' use of namespaces is organizational only and is not actually needed to disambiguate local names.

If you can make this simplifying assumption, then consider adjusting the xsd:imports basically to be xsd:includes and then using XSD flattening via

  1. the tool available in many commercial XML/XSD editors, or

  2. W. Paul Kiel's XML SchemaLightener, which has a flatten function, as a tool or as a starting point as it's written in XSLT

Your XSLT could then take the single, flattened XSD as input rather than the collection of separate XSDs.

See also

Upvotes: 1

Related Questions