Can an XML Schema sequence contain two element children named identically?-CodePudding

Here is a fragment of the XSD that I am trying to parse:

<xsd:complexType name="breakdownElementStructureRef">
    <xsd:sequence minOccurs="0">
        <xsd:element name="beUsageRef" type="breakdownElementUsageInBreakdownRef"/>
        <xsd:element name="beUsageRef" type="breakdownElementUsageInBreakdownRef"/>
    </xsd:sequence>
    <xsd:attribute name="uidRef" use="optional">
        <xsd:simpleType>
            <xsd:restriction base="uidRef">
                <xsd:pattern value="bes[1-9][0-9]*"/>
            </xsd:restriction>
        </xsd:simpleType>
    </xsd:attribute>
    <xsd:attribute name="uriRef" type="xsd:anyURI" use="optional"/>
    <xsd:assert test="(@uidRef and not(beUsageRef) and not(@uriRef)) or (not(@uidRef) and beUsageRef and not(@uriRef)) or (not(@uidRef) and not(beUsageRef) and @uriRef)"/>
</xsd:complexType>

For those who want more context, this is a fragment from this full schema.

Specifically my question relates to the xsd:element name="beUsageRef"... that is duplicated.

I am expecting this to be illegal XSD, but I could be wrong which is why I am looking for a definitive - yes it's legal, or no that is a mistake.

I have a possibly erroneous idea that each child of a complexType must be named, which is why I am expecting this to be in fact be illegal and a mistake in the XSD.

XML gurus out there, am I showing my relational DB bias and this is perfectly fine--a way to ensure that that the child repeats in multiples of 2 if at all?

I am specifically not asking about the semantic intent of the author of the XSD. I understand the xsd:assert specifies only one of beUsageRef, or uidRef or uriRef can exist, which implies a pattern of 2 "beUsageRef" children does not make sense since neither of the attributes can be duplicated, but I am not looking for that level of analysis.

I have been to the w3c and read the XSD spec. I couldn't parse the mathematical representations there to successfully answer my question -- I couldn't figure out if I should be trying to follow rules on a complexType, or on particles. So I seek the wisdom of my betters.

For this question, I just want to know if the fragment I posted is legal, valid, correct XSD? I just want to know if that fragment breaks an XSD rule of any kind.

CodePudding user response：

I think this is a valid schema, you can have the same element several times in a row as long as it is deterministic and not ambiguous. So the parser can find either those two beUsageRef with the given type or none of them, that is fine in my view.

CodePudding user response：

I am looking for a definitive - yes it's legal, or no that is a mistake.

Yes, it's legal.

this is perfectly fine--a way to ensure that that the child repeats in multiples of 2 if at all?

Almost right. Not multiples, because the sequence containing these two element refs is not allowed to repeat. The only two allowable options are

no occurrences of beUsageRef, because the enclosing sequence has minOccurs="0"
2 occurrences of beUsageRef, because if the first occurs in the XML then the enclosing sequence has occurred. Therefore the second beUsageRef must also occur (both element refs have minOccurs=1, implicitly).

I couldn't figure out if I should be trying to follow rules on a complexType, or on particles.

The element refs pointing to beUsageRef are particles because they are members of a sequence group.

I just want to know if that fragment breaks an XSD rule of any kind.

This may help with similar questions: https://www.utilities-online.info/xsdvalidation