[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Schema validation question

• To: Linda Grimaldi <grimlinda@e...>
• Subject: Re: Schema validation question
• From: Bob Foster <bob@o...>
• Date: Tue, 04 May 2004 23:19:31 -0500
• Cc: xml-dev@l...
• In-reply-to: <3196691.1083705512298.JavaMail.root@f...>
• References: <3196691.1083705512298.JavaMail.root@f...>
• User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.5) Gecko/20031007

This would be trivial to do in RELAX NG, but why do it? If the value of 
one element is fixed by the name of another, then the value is redundant 
and may be omitted. Such a thing is only useful for producing error 
messages.

Bob Foster

Linda Grimaldi wrote:
> I have been out of the XML world for a little while now, but someone just asked me a question and I have to admit, it's going to keep me up tonight if I can't come up with a better solution.  And I'm a little rusty, to boot...
> 
> The issue is one of data dependency and validation.  The goal is to enforce a restriction on pairs of data values- if the data associated with the node <Fruit><Name> is "Apple", then the value of the sibling node <Fruit><Abbrev> must be "APL".  Effectively, the goal is to enforce a specific mapping from one enum into another within an XML document. The only way my associate could figure to capture this dependency is as follows:
> 
> <xs:complexType name="Fruit">
> 	<xs:annotation>
> 		<xs:documentation>Defines the valid values for fruit types</xs:documentation>
> 	</xs:annotation>
> 	<xs:choice>
> 		<xs:element name="APL" type="xs:token" fixed="Apple"/>
> 		<xs:element name="ORG" type="xs:token" fixed="Orange"/>
> 		<xs:element name="BAN" type="xs:token" fixed="Banana"/>
> 		<xs:element name="LEM" type="xs:token" fixed="Lemon"/>
> 		<xs:element name="GRP" type="xs:token" fixed="Grape"/>
> 	</xs:choice>
> </xs:complexType>
> 
> All in all, this strikes me as a very ugly, even offensive, solution.  My initial response was to forget about the validation- it wasn't worth it.  Is there a way to preserve the data dependency validation without warping the instance document as the above schema fragment would require?
> 
> Thanks,
> Linda

• Follow-Ups:
  • Re: Schema validation question
    • From: "Chiusano Joseph" <chiusano_joseph@b...>

• References:
  • Schema validation question
    • From: Linda Grimaldi <grimlinda@e...>

• Prev by Date: Re: Schema validation question
• Next by Date: RE: You call that a standard?
• Previous by thread: Re: Schema validation question
• Next by thread: Re: Schema validation question
• Index(es):
  • Date
  • Thread