RE: [xml-dev] generate common xml shema from multiple xml instances

From
<>
To
Date
2009-06-17T11:44:45Z
ID
<>
Thread
RE: [xml-dev] generate common xml shema from multiple xml instances
> From: Rita Shen [mailto:]

> Hi,
>
> How can I generate a common XML schema for multiple xml instances??
> I know many tools can generate an XSD from a single XML instance.
> But in my case, I want to get the unknown common XSD for a bunch of XML
> instances.

Probably not what you want, but it is pretty easy to generate a "usage
schema" for Schematron.

You

  1) extract every simple XPath in all the documents, one per line, eg
       /book
       /book/section
       /book/section/heading
  2) run this through sort -ur     (if you are not running in UNIX, you can
    download the free GNU utility)  to get a reverse unique sort of the
lines.
  3) for each line, generate a Schematron rule with the line as the
context and an  <assert context="true()" /> assertion and a fallthrough
case. Eg

<pattern>
   <rule context="">
     <assert test="/book/section/heading" />
     <assert test="/book/section" />
     <assert test="/book" />
     <assert test="*" />Only the elements in the test documents can be
    used.</assert>
   </rule>
</pattern>

There are other varieties of things you could test. For example just each
element with no ancestor paths, or just parent/child pairs.

Cheers
Rick Jelliffe