XML and databases

1

• XML and databases

2
XML and Databases

• Data today
• Structured - Info in databases
• Data organized into chunks, similar entities
  groups together
• Descriptions for entities in groups same
  format, length, etc.
• Semi-structured data has certain structure, but
  not all items identical
• Schema info may be mixed in with data values
• Similar entities grouped together may have
  different attributes
• Self-describing data
• May be displayed as a graph
• Unstructured data
• Data can be of any type, may have no format or
  sequence
• Web pages in HTML
• Video, sound, images

3
XML documents and DBMS

• Options
• Use a DBMS to store XML documents as text
• If DBMS has module for document processing
• Use DBMS to store XML document contents as data
  elements
• Works if all documents have same structure
• map XML schema to DB schema
• Design special DBMS for storing XML data
• New type of XML DBMS designed, e.g. based on
  hierarchical model (Tamino)
• Create XML documents from preexisting RDBS and
  store into DB

4
Why XML

• Want Data source database with Web interface
• Specify content and format of Web pages with HTML
• Use HTML tags (predefined) for formatting Web
  documents
• HTML not suitable for specifying structured data
  from databases
• Does not contain schema information
• only how to display information
• XML standard for structuring and exchanging
  data over Web

5
Types of XML documents

• Data-centric documents have small data items
  following a specific structure
• Document-centric documents with large amount of
  text (little structured data)
• Hybrid documents with structured and
  unstructured data

6
XML

• Basic object is XML document
• Structuring concepts
• Elements (tags)
• Attributes
• XML attributes describe properties and
  characteristics of elements (tags)
• Also
• Entities, identifiers, references

7
Elements

• Elements identified by
• Start tag lt gt
• End tag lt/gt
• Simple elements data values
• Complex elements constructed from other
  elements hierarchically
• XML called a tree or hierarchical model
• No limit on number of nesting elements
• See fig. 27.1

8
Schemaless

• If semistructured data schemaless XML document
  and it is standalone
• lt?xml version 1.0 standaloneyes?gt
• (no corresponding file, e.g. DTD, specifying
  schema)

9
Well-formed

• XML document is well formed if
• Starts with XML declaration to indicate version
  and other relevant attributes
• Single root element
• Element matching pair of start/end tags within
  parent element
• Syntactically correct
• Can be processed to create internal tree

10
Next step?

• Want to use info in XML document to determine
  schema of database
• To do this, parse document to create tree
  structure of data
• After parsed, if not stand alone, can compare to
  definition of data structure (DTD of XML schema)
  to validate

11
Parsing summary

• To specify structure of data of semi-structured
  data
• DTD, XML Schema
• Each time access data, must parse document to
  create tree structure of data
• After parsed, compare to definition of data
  structure (DTD or XML schema) to validate
• Parsing SLOWS down the process

12
Parsing create internal tree

• Whole document must be parsed beforehand to
  generate tree
• Set of API functions to manipulate tree and
  parsing models
• DOM (Document Object Model) - uses main memory to
  parse entire document
• SAX allows processing XML documents on the fly
  (also good for streaming XML documents)
• Once parsed, allows validation of XML documents
  against
• DTD (Document Type Definition) file or
• XML schema file
• Valid means well formed and elements must follow
  structure and types specified in the separate
  schemas

13
XML DTD (Fig. 27.2)

• First specify root tag
• Parenthesis following element can be
• Type
• names of other elements (children)
• If PCDATA, means element is a leaf node (parsed
  character data string)
• Parenthesis can be nested
• indicates either

14
DTD contd

• To check for conformance to DTD add to XML
  document
• lt?xml version1.0 standalonenogt
• lt!DOCTYPE project SYSTEM proj.dtdgt
• Could also include DTD doc at beginning of XML
  doc
• Problems with DTD
• datatypes not general
• Special syntax requires specialized processors
• Elements must follow ordering of document

15
DTD - Notation for specifying elements

• - element can be repeated 0 or more times
• - element can be repreated 1 or more times
• ? element can be repeated zero or one times
• If no symbol, element must appear exactly once
• Type is specified in parenthesis (PCDATA)
  means parsed character data

16
XML Schema (Fig. 27.3)

• Alternative to (evolution from) DTD
• Standard for specifying structure of XML
  documents
• xsd XML schema definition

17
XML Schema

• Same syntax rules, so same processors on both
• Could display the entire Company database as a
  single document
• Could store DB in XML format instead of
  relational DB

18
Features of XML Schema

• 1) To identify XML schema language elements used,
  specify a file at a Web site location
• Each such definition is XML namespace
• ltxsdschema xmlnsxsdhttp//www.w3.org/2001/XML
  Schemagt
• File name assigned to xsd, and this variable used
  as prefix to all XML schema commands

19
Features contd

• 2) Annotation, documents and language
• Used for providing comments and other
  descriptions, e.g. en means english
• 3) Elements and types
• xsdelement - specifies element name
• xsdComplexType if elements and children
• xsdsequence ordered set of element types,
  e.g. dept, employee, etc.,

20
Features contd

• 4) First-level elements
• specified in element tags
• Element type, minimum and maximum occurrences
• MinOccurs, etc.
• also
• xsdkey - PK
• xsdunique tag, but must give constraint a name
• xsdkeyref foreign keys

21
Features contd

• Structures of complex elements - complex types
• Composite (compound) attributes complex types
• http//www.deepx.com/resources/quickref/XSD1-1.0.p
  df

22
XPath Language

• Addresses parts of an XML document
• Language mainly consists of location paths and
  expressions
• Provide a common syntax and semantics for
  functionality shared between Extensible
  Stylesheet Language Transformations (XSLT)
• XSLT - Language to create new document to
  transform the format of XML data into data of
  other formats (human readable) - eg. XML data
  into HTML, plain text, PDF
• XSLT - Describes how files encoded in XML are to
  be formatted or transformed
• XSLT functional language - no states, domain
  specific

23
To Query

• XPointer
• Specify position in XML document so other
  documents can link to it
• XPath
• facilities for manipulation of strings, numbers
  and booleans

24
To Query

• Use XPath (XML Path Language) to retrieve data
• XPath expression language, based on tree
  representation of XML document
• Small query language
• Provides ability to navigate around the tree
• Addresses specific parts of XML document
• Common syntax and behaviour model between
  Xpointer (addresses components of XML based
  internet media) and XSLT (translates XML into
  human readable documents)

25
XPath location path and expressions

• A location path is e.g. childparaposition(1)
  
• XPath expressions
• Returns collection of element nodes that satisfy
  patterns specified in expression
• Name with qualifier conditions
• Separators
• / means tag must appear as child of previous
  parent tag
• // means tag can appear as descendant of previous
  tag at any level

26
XPath

• To access whole XML document
• Doc(www.company.com/info.xml)/company
• /company/department
• //employee employeeSalary gt 70000/employeeName
• /company/employee employeeSalary gt
  70000/employeeName
• /company/project/projectWorker hours ge 20.0

27
To Query

• XQuery (like SQL) to query data using XPath
  expressions
• Based on SQL-like FLOWR (FLWOR) for joins
• For, Let, Where, Orderby Return

28
XQuery querying in XML

• For ltvariable bindings to individual nodesgt
• LET ltvariable binding to collections of nodesgt
• WHERE ltqualifer conditionsgt
• RETURN ltquery result specificationgt

29
Example query (Fig. 26.15)

• for x IN
• doc(www.company.com/info.xml)/company/employee
• WHERE x/employeeSalary gt 70000
• RETURN ltresgt x/employeeName/firstName,
• x/employeeName/lastName lt/resgt
• What does this do?
• What do you think of XQuery?

30
Research Topics

• XML to relational data mapping
• Updating XML in RDBMS
• XML and access control
• XML Parsing
• Some references on this topic
• XA Brief History of XML
• XML parsing A Threat to DB Performance
• XML Screamer

31
Oracle and XML

• /Schema for XML when using SQLPlus
• //A DTD is not needed!!!
• Create table Company of XMLType
• The rest of the definition
• The same fields do not have to be specified in
  every element
• Use set long 500
• to show all values in table

32
SELECT extractValue(OBJECT_VALUE,
'/Company/Employee/Fname') FNAME From Company
33
SELECT extractValue(OBJECT_VALUE,
'/Company/Employee/Fname') FNAME FROM Company
WHERE existsNode(OBJECT_VALUE, '/Company/Employee/
Salary lt"40000"') 1