Home > Articles > Web Services > XML

  • Print
  • + Share This
This chapter is from the book

This chapter is from the book

Simple Type Definitions

Named Simple Types

Simple types can be either named or anonymous. Named simple types are always defined globally (i.e., their parent is always schema or redefine) and are required to have a name that is unique among the data types (both simple and complex) in the schema. The XSDL syntax for a named simple type definition is shown in Table 1.

The name of a simple type must be an XML non-colonized name, which means that it must start with a letter or underscore, and may only contain letters, digits, underscores, hyphens, and periods. You cannot include a namespace prefix when defining the type; it takes its namespace from the target namespace of the schema document.

All of the examples of named types in this book have the word "Type" at the end of their names, to clearly distinguish them from element-type names and attribute names. However, this is not a requirement; you may in fact have a data type definition and an element declaration using the same name.

Example 1 shows the definition of a named simple type Dress-SizeType, along with an element declaration that references it. Named types can be used in multiple element and attribute declarations.

Table 1 XSDL Syntax: Named Simple Type Definition

Name

simpleType

Parents

schema, redefine

Attribute name

Type

Required/default

Description

id

ID

unique ID

name

NCName

required

simple type name

final

"#all" | list of ("extension" |

"restriction" |

"list" | "union")

defaults to finalDefault of schema

whether other types can be derived from this one, see section "Preventing Simple Type Derivation"

Content

annotation?, (restriction | list | union)


Example 1 Defining and Referencing a Named Simple Type

<xsd:simpleType name="DressSizeType">
 <xsd:restriction base="xsd:integer">
  <xsd:minInclusive value="2"/>
  <xsd:maxInclusive value="18"/> </xsd:restriction>
 </xsd:restriction>
</xsd:simpleType> 
<xsd:element name="size" type="DressSizeType"/> 

Anonymous Simple Types

Anonymous types, on the other hand, must not have names. They are always defined entirely within an element or attribute declaration, and may only be used once, by that declaration. Defining a type anonymously prevents it from ever being restricted, used in a list or union, or redefined. The XSDL syntax to define an anonymous simple type is shown in Table 2.

Table 2 XSDL Syntax: Anonymous Simple Type Definition

Name

 

 

 

simpleType

 

 

 

Parents

 

 

 

element, attribute, restriction, list, union

 

 

 

Attribute name

Type

Required/default

Description

id

ID

 

unique ID

Content

 

 

 

annotation?, (restriction | list | union)

 

 

 


Example 2 shows the definition of an anonymous simple type within an element declaration.

Example 2 Defining an Anonymous Simple Type

<xsd:element name="size">
 <xsd:simpleType>
  <xsd:restriction base="xsd:integer">
   <xsd:minInclusive value="2"/>
   <xsd:maxInclusive value="18"/> 
  </xsd:restriction> 
 </xsd:simpleType>
</xsd:element> 

Design Hint: Should I Use Named or Anonymous Types?

The advantage of named types is that they may be defined once and used many times. For example, you may define a type named Product-CodeType that lists all of the valid product codes in your organization.

This type can then be used in many element and attribute declarations in many schemas. This has the advantages of

  • encouraging consistency throughout the organization,

  • reducing the possibility of error,

  • requiring less time to define new schemas,

  • simplifying maintenance, because new product codes need only be added in one place.

Named types can also make the schema more readable, when the type definitions are complex.

An anonymous type, on the other hand, can be used only in the element or attribute declaration that contains it. It can never be redefined, have types derived from it, or be used in a list or union type. This can seriously limit its reusability, extensibility, and ability to change over time.

However, there are cases where anonymous types are preferable to named types. If the type is unlikely to ever be reused, the advantages listed above no longer apply. Also, there is such a thing as too much reuse. For example, if an element can contain the values 1 through 10, it does not make sense to try to define a data type named OneToTen-Type that is reused by other unrelated element declarations with the same value space. If the value space for one of the element declarations that uses the named data type changes, but the other element declarations do not change, it actually makes maintenance more difficult, because a new data type needs to be defined at that time.

In addition, anonymous types can be more readable when they are relatively simple. It is sometimes desirable to have the definition of the data type right there with the element or attribute declaration.

Every simple type is a restriction of another simple type, known as its base type. It is not possible to extend a simple type, except to add attributes, which results in a complex type. Every new simple type restricts the value space of its base type in some way. Example 3 shows a definition of DressSizeType that restricts the built-in type integer.

Example 3 Deriving a Simple Type from a Built-in Simple Type

<xsd:simpleType name="DressSizeType">
 <xsd:restriction base="xsd:integer">
  <xsd:minInclusive value="2"/>
  <xsd:maxInclusive value="18"/>
  <xsd:pattern value="\d{1,2}"/> 
 </xsd:restriction>
</xsd:simpleType> 

Simple types may also restrict user-derived simple types that are defined in the same schema document, or even in a different schema document. For example, you could further restrict DressSizeType by defining another simple type, MediumDressSizeType, as shown in Example 4.

Example 4 Deriving a Simple Type from a User-Derived Simple Type

<xsd:simpleType name="MediumDressSizeType">
 <xsd:restriction base="DressSizeType">
  <xsd:minInclusive value="8"/>
  <xsd:maxInclusive value="12"/> 
 </xsd:restriction>
</xsd:simpleType> 

A simple type restricts its base type by applying facets to restrict its values. In Example 4, the facets minInclusive and maxInclusive are used to restrict the value of MediumDressSizeType to be between 8 and 12 inclusive.

Defining a Restriction

The syntax for a restriction element is shown in Table 3. You must specify one base type either by using the base attribute, or by defining the simple type anonymously using a simpleType child. The alternative of using a simpleType child is generally only useful when restricting list types.

Table 3 XSDL Syntax: Simple Type Restriction

Name

restriction

Parents

simpleType

Attribute name

Type

Required/default

Description

id

ID

unique ID

base

QName

either a base attribute or a simpleType child is required

simple type name that is being restricted

Content

annotation? , simpleType? , (minExclusive | minInclusive | maxExclusive | maxInclusive | length | minLength | maxLength | totalDigits | fractionDigits | enumeration | pattern | whiteSpace)*


Within a restriction element, you can specify any of the facets, in any order. However, the only facets that may appear more than once in the same restriction are pattern and enumeration. It is legal to define a restriction that has no facets specified. In this case, the derived type allows the same values as the base type.

Overview of the Facets

The available facets are listed in Table 4.

The XSDL syntax for applying a facet is shown in Table 5. All facets must have a value attribute, which has different valid values depending on the facet. Most facets may also have a fixed attribute. Certain facets are not applicable to some types. For example, it does not make sense to apply the fractionDigits facet to a character string type. There is a defined set of applicable facets for each of the built-in types. (Technically, it is the primitive types that have applicable facets, with the rest of the built-in types inheriting that applicability from their base types. However, since most people do not have the built-in type hierarchy memorized, it is easier to list applicable facets for all the built-in types.)

If a facet is applicable to a built-in type, it is also applicable to atomic types that are derived from it. For example, since the length facet is applicable to string, if you derive a new type from string, the length facet is also applicable to your new type.

Table 4 Facets

Facet

Meaning

minExclusive

value must be greater than x

minInclusive

value must be greater than or equal to x

maxInclusive

value must be less than or equal to x

maxExclusive

value must be less than x

length

the length of the value must be equal to x

minLength

the length of the value must be greater than or equal to x

maxLength

the length of the value must be less than or equal to x

totalDigits

the number of significant digits must be less than or

 

equal to x

fractionDigits

the number of fractional digits must be less than or equal to x

whiteSpace

the schema processor should either preserve, replace, or collapse whitespace depending on x

enumeration

x is one of the valid values

pattern

x is one of the regular expressions that the value may match


Inheriting and Restricting Facets

When a simple type restricts its base type, it inherits all of the facets of its base type, its base type's base type, and so on back through its ancestors. Example 4 showed a simple type MediumDressSizeType whose base type is DressSizeType. DressSizeType has a pattern facet which restricts its value space to one or two-digit numbers. Because MediumDressSizeType inherits all of the facets from DressSizeType, this same pattern facet applies to MediumDressSizeType also. Example 5 shows an equivalent definition of MediumDressSize-Type, where it restricts integer and has the pattern facet applied.

Sometimes a simple type definition will include facets that are also specified for one of its ancestors. In Example 4, MediumDressSize-Type includes minInclusive and maxInclusive, which are also applied to its base type, DressSizeType. The minInclusive and maxInclusive facets of MediumDressSizeType (whose values are 8 and 12, respectively) override those of DressSizeType (2 and 18, respectively).

Table 5 XSDL Syntax: Facet

Name

minExclusive, minInclusive, maxExclusive, maxInclusive, length, minLength, maxLength, totalDigits, fractionDigits, enumeration, pattern, whiteSpace

Parents

restriction

Attribute name

Type

Required/default

Description

id

ID

unique ID

value

various

required

value of the restricting facet

fixed

boolean

false; n/a for pattern,

enumeration

whether the facet is fixed and therefore cannot be restricted further

Content

annotation?


Example 5 Effective Definition of MediumDressSizeType

<xsd:simpleType name="MediumDressSizeType">
 <xsd:restriction base="xsd:integer">
  <xsd:minInclusive value="8"/>
  <xsd:maxInclusive value="12"/>
  <xsd:pattern value="\d{1,2}"/> 
 </xsd:restriction>
</xsd:simpleType> 

It is a requirement that the facets of the derived type (in this case, MediumDressSizeType) be more restrictive than those of the base type. In Example 6, we define a new restriction of DressSizeType, called SmallDressSizeType, and set minInclusive to 0. This type definition is illegal, because it attempts to expand the value space by allowing 0, which was not valid for DressSizeType.

Example 6 Illegal Attempt to Extend a Simple Type

<xsd:simpleType name="SmallDressSizeType">
 <xsd:restriction base="DressSizeType">
  <xsd:minInclusive value="0"/>
  <xsd:maxInclusive value="6"/> 
 </xsd:restriction>
</xsd:simpleType> 

This rule also applies when you are restricting the built-in types. For example, the short data type has a maxInclusive value of 32767. It is illegal to define a restriction of short that sets maxInclusive to 32768.

Although enumeration facets can appear multiple times in the same type definition, they are treated in much the same way. If both a derived type and its ancestor have a set of enumeration facets, the values of the derived type must be a subset of the values of the ancestor. Likewise, the pattern facets specified in a derived type must allow a subset of the values allowed by the ancestor types. Schema processors will not necessarily check that the regular expressions represent a subset, but it will instead validate instances against the patterns of both the derived type and all the ancestor types, effectively taking the intersection of the pattern values.

Fixed Facets

When you define a simple type, you can fix one or more of the facets. This means that further restrictions of this type cannot change the value of the facet. Any of the facets may be fixed, with the exception of pattern and enumeration. Example 7 shows our DressSizeType with fixed minExclusive and maxInclusive facets, as indicated by a fixed attribute that is set to true.

Example 7 Fixed Facets

<xsd:simpleType name="DressSizeType">
 <xsd:restriction base="xsd:integer">
  <xsd:minInclusive value="2" fixed="true"/>
  <xsd:maxInclusive value="18" fixed="true"/>
  <xsd:pattern value="\d{1,2}"/> 
 </xsd:restriction>
</xsd:simpleType> 

With this definition of DressSizeType, it would have been illegal to define the MediumDressSizeType, as shown in Example 4, because it attempts to override the minInclusive and maxInclusive facets, which are now fixed. Some of the built-in types have fixed facets that cannot be overridden. For example, the built-in type integer has its fractionDigits facet fixed at 0, so it is illegal to derive a type from integer and specify a fractionDigits that is not 0.

Design hint: When Should I Fix a Facet?

Fixing facets makes your type less flexible, and discourages other schema authors from reusing it. Keep in mind that any types that may be derived from your type must be more restrictive, so you are not at risk that your type will be dramatically changed if its facets are unfixed.

A justification for fixing facets might be that changing that facet value would significantly alter the meaning of the type. For example, suppose you want to define a simple type that represents price. You define a Price type, and fix the fractionDigits at 2. This still allows other schema authors to restrict Price to define other types, such as, for example, a SalePrice type whose values must end in 99. However, they cannot modify the fractionDigits of the type, because this would result in a type not representing a price with both dollars and cents.

  • + Share This
  • 🔖 Save To Your Account