Updating XML

1

• Updating XML
• -Igor Tatarinov, Zachary G Ives, Alon Y.Halevy,
  Daniel S.Weld
• ACM SIGMOD 2001
• Presented By
• Deepanwita Roy

2
Outline

• Basic update operations
• XQuery
• Extensions to XQuery
• Storing XML in RDBMS
• Implementing update operations when XML is stored
  in relational database

3
XML

• Node labeled tree with references
• Types of XML Content
• Attribute
• IDREFs
• Element
• PCDATA

4
Outline

• Basic update operations
• XQuery
• Extentions to XQuery
• Implementing update operations when XML is stored
  in relational database

5
Update Operations

• Delete(child)
• Rename(child,name)
• Insert(content)
• InsertBefore(ref, content)
• InsertAfter(ref, content)
• Replace(child,content)

6
Outline

• Basic update operations
• XQuery
• Extensions to XQuery
• Storing XML in RDBMS
• Implementing update operations when XML is stored
  in relational database

7
XQuery(FLWR statements)

• FOR-LET-WHERE-RETURN
• FOR loops through each node in the sequence
• LETbinds variable to the entire sequence of node
• WHERE evaluating an expression
• RETURNreturn information

8
XQuery Example
Returns ltbookInfogt element for every book element
that has a publisher child of Addison-Wesley
9
Outline

• Basic update operations
• XQuery
• Extensions to XQuery
• Storing XML in RDBMS
• Implementing update operations when XML is stored
  in relational database

10
XQuery Extensions for Updates

• Extend XQuery with a FOR..LET..WHERE..UPDATE
• where update is defined as
• DELETE child
• RENAME child TO name
• INSERT contentBEFORE AFTER child
• REPLACE child WITH content
• FOR binding IN Xpath-subexpr,..
• WHERE predicate1,update

11
Example
12
Deletion
Allows dangling references
13
Insertion
14
Replacement
15
Combining operations with nested updates
16
Outline

• Basic update operations
• XQuery
• Extensions to XQuery
• Storing XML in RDBMS
• Implementing update operations when XML is stored
  in relational database

17
Storing XML in Relations

• Mapping XML into Relations
• XML Results as Outer Unions
• Access support Relations for Path Expression
  Evaluation

18
Mapping XML into Relations

• Shared Inlining method exploring the DTD

19
XML Results as Outer Unions

• Simulates Nested cursor approach
• Separate tuples in the same stream
• Output tuples sorted so that child data comes
  after parent data

For the query
20
(No Transcript)
21
Access Support Relations for Path Expression
Evaluation

• Indexes a specified pathextends to XML mapping

• Without ASR three joins select Part tuples with
  Number 123 then join with Item,Order,Customer
  relations.
• With ASR two joins Part and ASR relations and
  then result joined with Customer.

22
Outline

• Basic update operations
• XQuery
• Extensions to XQuery
• Storing XML in RDBMS
• Implementing update operations when XML is stored
  in relational database

23
XML Deletion Techniques

• Trigger-based
• Using per-tuple triggers
• Using per-statement triggers
• Cascading delete
• ASR-Based Delete(Access Support Relations)

24
Trigger based delete

• A trigger is a little rule that gets fired
  whenever a certain event occurs.
• Per-tuple trigger the deleted tuple is still
  accessible as is its id.The trigger should delete
  all tuples from child relations that link to the
  tuple through their parentId attribute.
• Per-statement trigger the trigger is fired
  after all relevant tuples have been deleted.

25
Cascading Delete

• First Delete relevant tuples from parent relation
  and then deleting orphaned tuples from their
  children.

Same performance as per-statement based trigger.
26
XML Insertion techniques

• Tuple-based Insert
• Table-based Insert
• ASR-Based Insert

27
Tuple based Insert

• Querying the source subtree and reading a tuple
  at a time.
• Every source element given a new identifier.
• The Sorted Outer Union Tuple partitioned back
  tables.
• Memory requirementslow
• Large penalty in terms of number of SQL inserts.

28
Table Based Insert

• Remap all source tuples in a single operation
• Insert remapped tuples within each underlying
  table
• Performance better because it executes a single
  SQL insertion per data relation

29
Conclusions

• Efficient storage of XML documents in RDBMS
• Updating XML documents and reflecting it in
  relational database.

30
Thank You.