XMLsBP4Db.md

XML Schema Best Practices for Databinding

empty elements

Most programming languages implement unitialized values as NULL, ZERO, random data, something special (JavaScript: undefined) or an empty string. In XML an element can be empty in two ways and in DOM the textContent property is an empty string for both.

Valid examples:

<TodoItem/>

<TodoItem></TodoItem>

So it should be interpreted as:

$TodoItem = "";

var TodoItem = "";

NULL: XsiNil

Most programming languages implement NULL values.

Best Practice:

• In the XSD: a value element may be marked as nillable (the default is false).
• In the XML: the value element marked as nillable must be marked as nil.

Definition example:

<xs:schema attributeFormDefault="unqualified" elementFormDefault="qualified" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element type="xs:string" nillable="true"/>
</xs:schema>

Valid example:

<TodoItem xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:nil="true"/>

Invalid example:

<TodoItem/>

So it should be interpreted as:

$TodoItem = null;

TodoItem = null;

Simple Lists: SequenceOfSimpleType

Most programming languages implement a mechanism to use simple types as a list. Simple types might also be called scalar types. Simple types are for example booleans, integers, strings and floats. Simple lists might be called arrays.

Best Practices:

• A list element should {must?} be a sequence containing one value element.
• A list element should {must?} not have attributes. To not interfere with the items.
• The value elements maxOccurs should be unbounded instead of the default 1. So the List can contain many items.
• The value elements minOccurs should be 0 instead of the default 1. So the List can contain zero items.
• The value elements nillable may be set. {JavaScript array gaps? PHP numerically-indexed array preserved?}

<xs:schema attributeFormDefault="unqualified" elementFormDefault="qualified" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element name="TodoList">
    <xs:complexType>
      <xs:sequence>
        <xs:element type="xs:string" name="TodoItem" maxOccurs="unbounded" minOccurs="0"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

Valid:

<TodoList>
</TodoList>

<TodoList>
  <TodoItem>Beer</TodoItem>
</TodoList>

<TodoList>
  <TodoItem>Eggs</TodoItem>
  <TodoItem>Milk</TodoItem>
</TodoList>

Invalid:

<TodoList>
  <TodoItem/>
</TodoList>

So it should be interpreted as:

$TodoList = []; // ZERO
$TodoList = ['Beer']; // ONE
$TodoList = ['Eggs', 'Milk']; // MANY

TodoList = []; // ZERO
TodoList = ["Beer"]; // ONE
TodoList = ["Eggs", "Milk"]; // MANY

Note how the element name TodoItem has no relevance to the code.

Complex lists: SequenceOfComplexType

Most programming languages implement a mechanism to use non-simple types as a list too. Complex types might also be called non-scalar types, structs, classes or anonymous objects. Complex lists might be called arrays too.

Best Practices:

• *all simple list best practices apply, please read these first
• The value elements type may be a globally defined complex type. To improve readability only.

<xs:schema attributeFormDefault="unqualified" elementFormDefault="qualified" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:complexType name="TodoItemType">
    <xs:all>
      <xs:element type="xs:string" name="text"/>
      <xs:element type="xs:boolean" name="done" nillable="true"/>
    </xs:sequence>
  </xs:all>
  <xs:element name="TodoList">
    <xs:complexType>
      <xs:sequence>
        <xs:element name="TodoItem" type="TodoItemType" maxOccurs="unbounded" minOccurs="0"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

Valid:

<TodoList>
</TodoList>

<TodoList xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
  <TodoItem>
    <text>beer</text>
    <done xsi:nil="true">
  </TodoItem>
</TodoList>

<TodoList>
  <TodoItem>
    <text>Eggs</text>
    <done>true</done>
  </TodoItem>
  <TodoItem>
    <text>Milk</text>
    <done>false</done>
  </TodoItem>
</TodoList>

Invalid example:

<TodoList>
  <TodoItem>
    <text/>
  <TodoItem>
</TodoList>

This does not fail because the text element is empty - that is still a valid empty string. This does fail because the done element is missing, which is not allowed by the complex type.

So it should be interpreted as:

$TodoList = [];
$TodoList[0] = new stdClass();
$TodoList[0]->text = "Beer";
$TodoList[0]->done = null;

TodoList = {
  "text": "Beer",
  "done": null
};

Dictionary / hash map

Many programming languages implement a special type for key and value list. In XML Schema one might define a complex list for this.

SECTION UNDER CONSTRUCTION :: THIS DOES NOT MAKE SENSE YET

Best Practice:

• The value item should use an attribute of the simple type string that is required.

<xs:schema attributeFormDefault="unqualified" elementFormDefault="qualified" xmlns:xs="http://www.w3.org/2001/XMLSchema">
  <xs:element name="TodoList">
    <xs:complexType>
      <xs:sequence>
        <xs:element type="xs:string" name="TodoItem" maxOccurs="unbounded" minOccurs="0"/>
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

Valid example:

<TodoList>
  <TodoItem identifier="45676773">
    <text>Eggs</text>
    <done>true</done>
  </TodoItem>
  <TodoItem identifier="4d696c6b">
    <text>Milk</text>
    <done>false</done>
  </TodoItem>
</TodoList>

So it should be interpreted as:

$TodoList = [];
$TodoList[0] = new stdClass();
$TodoList[0]->identifier = ''
$TodoList[0]->text = "Beer";
$TodoList[0]->done = null;
$TodoList[1] = new stdClass();
$TodoList[1]->identifier = ''
$TodoList[1]->text = "Beer";
$TodoList['4d696c6b']->done = null;

TodoList = {
  "text": "Beer",
  "done": null
};