BD11 Data Model

Type System

edge vs node labeling: labels are on the edges(JSON)/nodes (XML)

Shared properties

• distinction between atomic types and structured types
• same categories of atomic types
• lists and maps as structured types
• sequence type cardinalities: *(zero or more, repeated),?(zero or one, optional),+(one or more)

  Atomic types

• string
• number
  • arbitrary precision decimales and integers
  • signed and unsigned integer types
  • IEEE 754 standard: float: 32bits, ca 7digits, 10^-37 to 10^37; double: 64bits, ca 15digits, 10^-307 to 10^308
• booleans
• dates and times
  • dates: year+month+day
  • times: hours+minutes+seconds
  • timestamp: date+time
• Time interval
• binaries
• null

  Structured types

• associative arrays( maps) : JSON object, set of XML attributes
• ordered lists: JSON array, XML element

Validation

pipeline: document->well-formedness->validation

PSVI: post-schema-validation infoset: infoset + types

DTD validation

Element type declaration

• empty content: <!ELEMENT name EMPTY>
• simple content: only includes text, using #PCDATA, <!ELEMENT name (#PCDATA)>
• complex content: meaning including other elements: <!ELEMENT name (sub1/+/*/?,sub2, ...)>, the more advanced:<!ELEMENT name (bar*|(foo|foobar)+)? >
• mixed content: includes both other elements and text, e.g. <!ELEMENT name (#PCDATA|foo)*>, or can also use ANY <!ELEMENT name ANY>

  Attribute-List Declaration

  Format: <!ATTLIST ele-name attr-name ATTR-FORMAT CARDINALITY>
• attribute format: CDATA, NMTOKEN(s), ID(REFS)
• cardinality:
  • #REQUIRED: must have
  • #IMPLIED: can have
  • “string”: the value of the attribute, can be omitted.
  • #FIXED “string”: must written out, cannot be omitted.

XML Schema

Two files: .xsd: the schema defined file and .xml: the instance file

• .xsd: the definition of an element: <xs:element name="ele-name", type="xs:string/xs:integer/user-defined type..."/>
• .xml: within the defined element: using <ele-name xmlns:xsi="..." xsi:schema-location="schema-name.xsd">

Element types

Simply types: built-in

• strings: string, anyURI, QName;
• numbers: decimal, float, double, integer, …;
• booleans: boolean
• dates and times: dateTime, time, date, gYearMonth, gMonthDay, gYear, gMonth, gDay, dateTimeStamp
• time intervals: duration, yearMonthDuration, dayTimeDuration
• Binaries: hexBinary, base64Binary
• NUll

  User-defined types

  Simple types

• restriction: <xs:restriction base="built-in type"> <xs:length value="3">...</xs:restriction>
• list: <xs:list itemType="xs:string"/>, instance: <foo>foo bar foobar</foo>
• union: <xs:union memberTypes="xs:integer xs:boolean"/>, instance: <foo>true<foo>

  Complex types

• empty: <xs:complexType name="emptyType">
• simple content ??? what is different with the simple type?:
• complex content: can include sub-elements
• mixed conent: can include both text and sub-elements. <xs:complexType, name="..." mixed="true">

  Keys

  XML allows you to make an attribute as a key of the element. the definition is
  <xs:element ...>....<xs:key name="key-name"> <xs:selector xpath="the-aim-element"/><xs:field xpath="@the-key-attribute"/></xs:key></xs:element>

How to use? 1. simple type of attributes; 2. named types; 3. anonymous types: defined where to use, without being named.

Schema Location

without namespace

in the instance file: <xsi:noNamespaceSchemaLocation="schema.xsd">. Correspondingly, in the schema file, there is no definition on the namespace.

With namespace

in the instance file: <xsi:SchemaLocation="namespace/schema.xsd">.
In the schema file, <xs:schema xmlns:xs="..." targetNamespace="namespace" ...>

Import Schema

in the .xsd file, we can use types defined in other schema files, by using <xs:import namespace="other-schema-namespace" schemaLocation="other-schema-name.xsd"/>. To import, we use ref, in the sentence: <xs:element ref="prefix:defined-type" ...>

Other technologies

Avro: It is a language that allows you to define schema, an apache language.
Process:

1. read the language-independent schema
2. compilation the schema and put it in somewhere of the code
3. read in the data according to the schema.

XDM: XPath and XQuery Data Model, JDM: JSONiq Data Model

XDM: seven kinds of XML Nodes: Document Node, Element node, Attribute node, Text Node, Comment node, Processing instruction node, namespace node.

Abbreviation

PSVI: post-schema-validation infoset
DTD: document type declaration
XDM: XPath and XQuery Data Model
JDM: JSONiq Data Model

Further Reading

XML in a Nutshell Chapter 3, 17

Thanks Ruben and Christina for reading notes