«

Hello,
I'm learning W3C Schema 1.0 and have encountered a problem. I want to
make a complex type which requires to contain exactly 1 instance of each
of a list of elements (a, b, c) and at least once a choice of several
other elements ( d1 | d2 | d3 ). This in any order.

I have come up with the following which is (unfortunately) invalid XSD 1.0:
<complexType name="root">
<all>
<element name="a" type="string"></element>
<element name="b" type="string"></element>
<element name="c" type="string"></element>
<choice maxOccurs="unbounded" minOccurs="1">
<element name="d1" type="string"></element>
<element name="d2" type="string"></element>
<element name="d3" type="string"></element>
</choice>
</all>
</complexType>

If somebody can help me make up a valid schema for this, I would be very
thankful. Below are some XML examples to clarify what I expected.

Thanks Phil

=== XML example ===

VALID (a,b,c are there, and at least one d*)
<root>
<a />
<b />
<c />
<d1 />
</root>

VALID (order changed, that's OK)
<root>
<a />
<d1 />
<c />
<b />
</root>

VALID (multiple d* are OK)
<root>
<a />
<d2 />
<b />
<c />
<d3 />
<d2 />
</root>

INVALID (a appears twice, bad)
<root>
<a />
<b />
<c />
<d1 />
<a />
</root>

INVALID (no d* element, bad)
<root>
<a />
<b />
<c />
</root>

Philipp <> wrote in
<>:
> I'm learning W3C Schema 1.0 and have encountered a
> problem. I want to make a complex type which requires to
> contain exactly 1 instance of each of a list of elements
> (a, b, c) and at least once a choice of several other
> elements ( d1 | d2 | d3 ). This in any order.
>
> I have come up with the following which is (unfortunately)
> invalid XSD 1.0: <complexType name="root">
> <all>
> <element name="a" type="string"></element>
> <element name="b" type="string"></element>
> <element name="c" type="string"></element>
> <choice maxOccurs="unbounded" minOccurs="1">
> <element name="d1" type="string"></element>
> <element name="d2" type="string"></element>
> <element name="d3" type="string"></element>
> </choice>
> </all>
> </complexType>

W3C's schema definition language was not designed for
defining arbitrarily complex grammars. If your document
format is well-structured, expressing it in XML Schema
should be easy. If it's not, try using a more powerful
schema definition language or check validity on the
application side.

--
"...a Netscape engineer who shan't be named once passed a
pointer to JavaScript, stored it as a string and later
passed it back to C, killing 30..." --Blake Ross

Pavel Lepin wrote:
> Philipp <> wrote in
> <>:
>> I'm learning W3C Schema 1.0 and have encountered a
>> problem. I want to make a complex type which requires to
>> contain exactly 1 instance of each of a list of elements
>> (a, b, c) and at least once a choice of several other
>> elements ( d1 | d2 | d3 ). This in any order.
>>
>> I have come up with the following which is (unfortunately)
>> invalid XSD 1.0: <complexType name="root">
>> <all>
>> <element name="a" type="string"></element>
>> <element name="b" type="string"></element>
>> <element name="c" type="string"></element>
>> <choice maxOccurs="unbounded" minOccurs="1">
>> <element name="d1" type="string"></element>
>> <element name="d2" type="string"></element>
>> <element name="d3" type="string"></element>
>> </choice>
>> </all>
>> </complexType>
>
> W3C's schema definition language was not designed for
> defining arbitrarily complex grammars. If your document
> format is well-structured, expressing it in XML Schema
> should be easy. If it's not, try using a more powerful
> schema definition language or check validity on the
> application side.
>

After some reading on the web, I think my biggest problem is that my
elements are non-ordered. This makes sense for me both from the model
perspective (I don't care if eg. my Book object sets the Author or the
Title first) as well as from my implementation perspective (these values
are marshalled from a java Properties object which does not guaratee a
constant traversal order for its iterator.

Hopefully Xerces for Schema 1.1 will soon be available (and also
that Schema 1.1 fixes these shortcommings)

Phil