SemiStructured Data and XML

1
Semi-Structured Data and XML
2
Agenda

• Semi-Structured Data
• XML

3
Semi-Structured Data an Introduction

• What is structured data
• What is non-structured data
• What is semi-structured data
• How is semi-structured data represented?
• What can we do with semi-structured data?

4
What is Structured Data?

• Strongly typed variables/attributes
• (ie. int, float, string20)
• Every attribute in a relation is defined for all
  records
• Data is represented in some organized fashion

5
An Example of Structured Data
A relational database can be considered
structured data
6
What is Non-Structured Data?

• Data that has no type definitions
• Data is not organized according to any pattern
• No concept of variables or attributes

7
An Example of Non-Structured Data
Bob was born sometime in August of 1949. He has
a reasonable salary of 52000. Someone else was
born on the 12th of a different month, his name
is Bill. By the way, Bob was born on the 13th of
August.
As you can see, such data would be almost
impossible to have a computer automatically parse.
8
Then what is Semi-Structured Data?

• Anything in between structured and non-structured
  data!

9
Then what is Semi-Structured Data?

• Everything in between structured and
  non-structured data
• Variables are loosely typed
• x1 is valid, so is xhello
• A record does not need to have all attributes
  defined
• ie. In a database of cars, if we dont know the
  engine type, we can choose not to define the
  field for tha particular record. Whereas in a
  structured database, the attribute would be
  defined, but set to NULL.
• An attribute of a record could be another record
• It does not necessarily have to differentiate
  between an identifier and a value

10
So how is semi-structured data represented?

• Semi-Structured data can be represented as a tree

11
So how is semi-structured data represented?

• Semi-Structured data can be represented in the
  form of indented text

Bob Birthday 1949 August 13 Salary 52,0
00 Bill Birthday 1967 April
12
So how is semi-structured data represented?

• Semi-Structured data can be represented as a
  markup language (ie. HTML, XML, LISP, AceDB,
  Tsimmis)

ltemployee id3gt ltnamegtBoblt/namegt ltextensiongt55
13lt/extensiongt ltdepartmentgtSaleslt/departmentgt lts
alarygt45000lt/salarygt lt/employeegt ltemployee
id1gt ltnamegtEdlt/namegt ltextensiongt6766lt/extensi
ongt ltofficegt312lt/officegt ltdepartmentgtExecutivelt/
departmentgt ltsalarygtConfidentiallt/salarygt ltemploy
eegt
13
Overview

• Semi-Structured data is not necessarily created
  with the intention of being processed.
• ie. Web pages are not necessarily intended to be
  queried by a language like SQL the web designer,
  not taking this into consideration may not make
  it easy for the data to be processed by a machine.

14
What can we do with Semi-Structured Data?

• Since there is some structure, it can be scanned
  and parsed
• Once the data is parsed, we can query it using
  specialized query languages such as UnQL, GEXT
  and Lorel
• We can clean it up to be placed into a
  structured relational database

15
XML an Introduction to XML

• What is XML?
• What does it offer to creators of DBs?
• How can XML be used as a DB?
• Representations of XML
• Other features of XML
• Disadvantages to XML

16
Summary / Key Points of Semi-Structured data

• In between structured and non-structured data
• Loosely typed attributes
• Not all attributes need to be defined for every
  record
• Can be parsed and queried

17
What is XML?

• XML stands for eXtensible Markup Language
• Based on tags similar to HTML
• Actually, XHTML is a form of XML
• Used to define markup languages

18
What does XML offer to database designers?

• Readable by humans using Unicode or ASCII text
• Easy for computers to parse
• Can easily be used as back-end for web sites

19
How can XML be used as a database?
Consider the following data
ltemployee id3gt ltnamegtBoblt/namegt ltextensiongt55
13lt/extensiongt ltdepartmentgtSaleslt/departmentgt lts
alarygt45000lt/salarygt lt/employeegt ltemployee
id1gt ltnamegtEdlt/namegt ltextensiongt6766lt/extensi
ongt ltofficegt312lt/officegt ltdepartmentgtExecutivelt/
departmentgt ltsalarygtConfidentiallt/salarygt ltemploy
eegt
It can be written in XML as follows
Notice that this is semi-structured data, since
not all the fields are filled in and because they
are loosely typed.
20
In XML, there are few restrictions to how data
can be laid out

• The tag names can represent either attribute
  names or data itself
• Tag names can be defined to anything the creator
  wishes

21
But, there are still a few restrictions

• Every tag that is opened, must be closed.
• ltnamegtBoblt/namegt
• Close tag is not needed for empty data
• ltmyelement/gt
• If one tag is opened inside the field of another
  tag, it must be closed before the outer tag is
  closed.
• ltemployeegtltnamegtBoblt/employeegtlt/namegt
• ltemployeegtltnamegtBobgtlt/namegtlt/employeegt
• Tags are case sensitive

22
How can XML be represented?

• As a tree structure
• As text/markup tags

23
How can XML be represented?

• As a tree structure

Take our previous example

• Leaf nodes generally, but do not necessarily
  store the data
• Recent web browsers will show this structure

24
How can XML be represented?

• As a text/markup language

Take our previous example
ltemployee id3gt ltnamegtBoblt/namegt ltextensiongt55
13lt/extensiongt ltdepartmentgtSaleslt/departmentgt lts
alarygt45000lt/salarygt lt/employeegt ltemployee
id1gt ltnamegtEdlt/namegt ltextensiongt6766lt/extensi
ongt ltofficegt312lt/officegt ltdepartmentgtExecutivelt/
departmentgt ltsalarygtConfidentiallt/salarygt ltemploy
eegt
25
Other features of XML

• It is easy to parse
• It can be queried like a database
• It can be used with XSL Templates to easily
  generate web pages from data
• It can be used with DTS (Document Type
  Definition) to run as a fully structured database

26
Disadvantages to XML

• Difficult create indexes on
• Difficult to optimize queries
• Requires additional disk space
• Text format
• Redundant data in tags
• No single standard of how data should be stored
  in XML

27
Summary / Key points of XML

• Data stored using text-based markup language
• Can also be represented in tree format
• Can store structured and semi-structured data
• Easy to parse and query, but inefficient

28
Where to Get More Information

• Search the web, youll find something!