XPath

1
XPath
2
XPath

• identifies parts of an XML document
• XPath is not written in XML
• resembles (Unix) directory and filename path
  structures
• also represents simple data structures
• Numbers
• Booleans
• Strings

3
XPath node types

• root node
• denoted /
• absolute XPath expressions derive from /
• NOT the same as the document root element
• element nodes
• one for each type of XML element in a document
• nesting gives the usual parent-child tree
  structure of elements

4
XPath node types

• text nodes
• CDATA content of an element
• entity references resolved by the parser before
  node is resolved
• attribute nodes
• element nodes are parents of their attribute
  nodes
• attribute nodes are NOT children of their parents!

5
XPath node types

• comment nodes
• one for each comment in the XML document
• processing instruction nodes
• one for each processing instruction
• namespace nodes
• attached to each element or attribute node for
  which a namespace is specified

6
XPath node types

• not included as node types
• CDATA sections
• simply merged into text sections
• entity references
• e.g. lt
• simply merged in as literal strings
• Document Type Declarations
• default attributes defined by DTDs merged into
  attribute nodes by the parser

7
ltxml version 1.0gt lt?xml-stylesheet type
application/xml href people.xslgt lt?DOCTYPE
people lt!ELEMENT person (name, profession,
homepage?) gt lt!ATTLIST person born CDATA
IMPLIEDgt lt!ATTLIST person died CDATA
IMPLIEDgt ... gt
8
ltpeoplegt ltperson born1912 died1954gt ltnamegtltf
irst_namegtAlanlt/first-namegt ltlast_namegtTuringlt/la
st_namegt lt/namegt lt!-- Did the word computer
scientist exist in Turings day?
--gt ltprofessiongtcomputer scientistlt/professiongt ltp
rofessiongtmathematicianlt/professiongt ltprofessiongtc
ryptographerlt/professiongt lthomepage
xlinkhrefhttpwww.Turing.org.uk//gt lt/persongt
9
ltperson born1918 died1988gt ltnamegtltfirst_name
gtRichardlt/first-namegt ltmiddle_initialgtx50lt/mid
dle_initialgt ltlast_namegtFeynmanlt/last_namegt lt/nam
egt ltprofessiongtphysicistlt/professiongt lthobbygtplayi
ng the bongoeslt/hobbygt lt/persongt lt/peoplegt
10
XPath Tree Structure
11
location paths

• identifies a set of nodes in the document
• may be empty
• may be of different types
• constructed from location steps
• each step is evaluated with respect to the
  current context node
• root location path
• /

12
child element location steps

• single element name
• identifies all child elements of the current
  context node with that name
• example
• profession
• identifies 3 nodes if the context is the first
  person node (Turing)
• identifies 1 node if the context is the second
  person node (Feynman)

13
the context node

• the node that is currently matched when the XPath
  expression is evaluated

ltxsltemplate match persongt ltpgt ltxslvalue-of
select name /gt lt/pgt lt/xsltemplategt
14
attribute location steps

• syntax
• _at_attribute_name

ltxsltemplate match persongt lttrgt lttdgtltxslvalu
e-of select name /gtlt/tdgt lttdgtltxslvalue-of
select _at_born /gtlt/tdgt lttdgtltxslvalue-of select
_at_died /gtlt/tdgt lt/trgt lt/xsltemplategt
15
other location steps

• comment()
• identifies all children that are comment nodes
• text()
• identifies all children that are text nodes
• processing-instruction()
• identifies all children that are processing
  instruction nodes
• parameters can be used to specify precisely which
  nodes are identified
• namespace nodes cannot be referenced by XPath

16
examples

• ltxsltemplate match comment()gt
• ltigtComment deletedlt/igt
• lt/xsltemplategt
• ltxsltemplate match processing-instruction(xml
  -stylesheet)gt
• lth1gta stylesheet is attachedlt/h1gt
• lt/xsltemplategt

17
wildcards

• match different element and node types at the
  same time
• identifies all elements in context
• node()
• identifies all nodes in context
• _at_
• identifies all attributes in context

18
multiple matches

• use to separate multiple location steps
• identifies all nodes in context matching at least
  one of the steps
• examples
• professionhobby
• first_namemiddle_initiallast_name
• idxlinktype
• _at_

19
compound location paths

• combine location steps with / to form compound
  location paths
• each step is relative to the last
• / identifies the root path
• absolute paths start with /
• relative paths depend from the context node
• /people/person/name/first_name
• person/_at_born

20
selecting from descendants

• // selects from the context node and all its
  descendants
• examples
• //name
• /people/person//name
• //_at_id
• person//_at_id

21
selecting the parent element

• .. identifies the parent of the current node
• examples
• //_at_id..
• identifies all elements in the document with an
  attribute named id
• //middle_initial/../first_name
• identifies all first_name elements that are
  siblings of middle_initial elements
• ltfirst_namegtRichardlt/first_namegt

22
selecting the context node

• . indicates the context node
• example
• ltxsltemplate match comment()gt
• ltspan class commentgt
• ltxslvalue-of select . /gt
• lt/spangt
• lt/xsltemplategt

23
predicates

• Boolean expressions used to provide more
  discrimination in identifying nodes
• may attach a predicate to each location step
• syntax
• locationSteppredicate
• relational operators
• lt, lt, , gt, gt, !
• Boolean operators
• and, or, not()

24
predicate examples
//profession.physicist . here means the
string content of the profession
element //personprofessionphysicist single
quotes can be used to avoid parse
errors //person_at_idp4567 ltxslapply-templates
select //person_at_bornlt1976 /gt lt, gt etc
have to be escaped in well-formed XML documents
25
general XPath expressions

• Numbers
• 3.141, 22, 3 - 4 17 div 6 mod 2
• Booleans
• true(),false(), 32.5 lt 76.2
• Strings
• this is a string value
• ltso is this /gt

26
XPath functions

• return Booleans, Numbers, Strings or Node-sets
• no void functions
• example
• ltxsltemplate match persongt
• Personltxslvalue-of selectposition()/gt
• ltxslvalue-of select name/gt
• lt/xsltemplate

27
function examples
ltxsltemplate match persongt Personltxslvalue-o
f selectposition()/gt ltxslvalue-of select
name/gt lt/xsltemplategt produces a list
Person1, name1, Person2, name2, from name
elements ltxslapply-templates selectnamestar
ts-with(last_name,T)/gt apply a template to
all name elements, the string content of whose
last-name element begins with the character T
28
Schedule Change

• Week commencing 15th October
• Weds 17th October BC6 1100-1300
• Double Lecture, different venue
• no change to Monday lecture
• No Friday Lab
• Week commencing 22nd October
• No lectures
• Thursday Lab only available
• Week commencing 29th October
• Weds 31st October BC6 1100-1300
• Double Lecture, different venue
• no Monday lecture