=========================================================================
LECTURE NOTES - XML
=========================================================================

If you are interested in reading about the XML standard, please visit:

  http://www.w3.org/XML/

It will keep you busy for days.  A very short web page linked to that
site ("XML in 10 points") provides a light summary.


XML = Extensible Markup Language

-> A (relatively) new standard for data representation and exchange on
   the internet

-> A document format: a superset of HTML, a subset of SGML (roughly)

-> XML is to data what Java is to programming


Like SGML and HTML, basic XML consists of three things:

(1) Tagged elements, which may be nested within one another
(2) Attributes on elements
(3) Text

In HTML, tags denote formatting: <TITLE>, <I>, <TABLE>, etc.

In XML, tags denote meaning of data: <STUDENT>, <BOOKTITLE>, etc.

(To format XML data, use XSL - the Extensible Stylesheet Language - to
translate XML to HTML.)


Well-formed XML
===============

A well-formed XML document is any XML document that follows the basic
rules: matched tags, unique attribute names, etc.


Ex: bookstore data

<?XML VERSION="1.0" STANDALONE="yes"?>
<BOOKSTORE>
   <BOOK ISBN="0-13-861337-0" Price="$50">
      <TITLE>A First Course in Database Systems</TITLE>
      <AUTHORS>
         <AUTHOR>
            <FIRSTNAME>Jeffrey</FIRSTNAME>
            <LASTNAME>Ullman</LASTNAME>
         </AUTHOR>
         <AUTHOR>
            <FIRSTNAME>Jennifer</FIRSTNAME>
            <LASTNAME>Widom</LASTNAME>
         </AUTHOR>
      </AUTHORS>
   </BOOK>
   <BOOK ISBN="0-13-040264-8" Price="$60">
      <TITLE>Database System Implementation</TITLE>
      <AUTHORS>
         <AUTHOR>
            <FIRSTNAME>Hector</FIRSTNAME>
            <LASTNAME>Garcia-Molina</LASTNAME>
         </AUTHOR>
         <AUTHOR>
            <FIRSTNAME>Jeffrey</FIRSTNAME>
            <LASTNAME>Ullman</LASTNAME>
         </AUTHOR>
         <AUTHOR>
            <FIRSTNAME>Jennifer</FIRSTNAME>
            <LASTNAME>Widom</LASTNAME>
         </AUTHOR>
      </AUTHORS>
      <REMARK>
       Buy this book bundled with "A First Course", it's a great deal!
      </REMARK>
   </BOOK>
</BOOKSTORE>


A well-formed XML document can contain regular data (as above) or very
irregular data.


Valid XML
=========

It is possible to define a "schema" for XML data, called a Document
Type Descriptor (DTD).

A DTD is a grammar that describes the legal nesting of tags and
attributes.

<!DOCTYPE Bookstore [
   <!ELEMENT BOOKSTORE (BOOK | MAGAZINE)*>
   <!ELEMENT BOOK (TITLE, AUTHORS, REMARK?)>
   <!ATTLIST BOOK ISBN CDATA Price CDATA>
   <!ELEMENT MAGAZINE (TITLE)>
   <!ATTLIST MAGAZINE Month CDATA Year CDATA>
   <!ELEMENT TITLE (#PCDATA)>
   <!ELEMENT AUTHORS (AUTHOR+)>
   <!ELEMENT REMARK (#PCDATA)>
   <!ELEMENT AUTHOR (FIRSTNAME, LASTNAME)>
   <!ELEMENT FIRSTNAME (#PCDATA)>
   <!ELEMENT LASTNAME (#PCDATA)>
]>


The DTD is specified at the top of the document or in a separate file
referenced at the top of the document.  In both cases use
STANDALONE="no".

Q: What are the benefits of using DTDs?










ID and IDREF(S) Attributes
==========================

Element pointers: assign a special "ID" attribute to an element, then
point to that element with a special "IDREF(S)" attribute in another
element.


Ex: reorganized bookstore

<?XML VERSION="1.0" STANDALONE="no"?>
<BOOKSTORE>
   <BOOK ISBN="0-13-861337-0" Price="$50" Authors="JU JW">
      <TITLE>A First Course in Database Systems</TITLE>
   </BOOK>
   <BOOK ISBN="0-13-040264-8" Price="$60" Authors="HG JU JW">
      <TITLE>Database System Implementation</TITLE>
      <REMARK>
         Buy this book bundled with
         <BOOKREF book="0-13-861337-0"></BOOKREF>
         It's a great deal!
      </REMARK>
   </BOOK>
   <AUTHOR Ident="HG">
      <FIRSTNAME>Hector</FIRSTNAME>
      <LASTNAME>Garcia-Molina</LASTNAME>
   </AUTHOR>
   <AUTHOR Ident="JU">
      <FIRSTNAME>Jeffrey</FIRSTNAME>
      <LASTNAME>Ullman</LASTNAME>
   </AUTHOR>
   <AUTHOR Ident="JW">
      <FIRSTNAME>Jennifer</FIRSTNAME>
      <LASTNAME>Widom</LASTNAME>
   </AUTHOR>
</BOOKSTORE>


DTD for this data:

<!DOCTYPE Bookstore [
   <!ELEMENT BOOKSTORE (BOOK*, AUTHOR*)>
   <!ELEMENT BOOK (TITLE, REMARK?)>
   <!ATTLIST BOOK ISBN ID Price CDATA Authors IDREFS>
   <!ELEMENT TITLE (#PCDATA)>
   <!ELEMENT REMARK (#PCDATA | BOOKREF)+)>
   <!ELEMENT BOOKREF>
   <!ATTLIST BOOKREF book IDREF>
   <!ELEMENT AUTHOR (FIRSTNAME, LASTNAME)>
   <!ATTLIST AUTHOR Ident ID>
   <!ELEMENT FIRSTNAME (#PCDATA)>
   <!ELEMENT LASTNAME (#PCDATA)>
]>


Querying XML
============

-> XML turns the Web into one big database

Several languages have been proposed for querying XML data.

  - We developed one at Stanford called "Lorel", based on OQL.

  - There is a recently proposed standard called XQuery.

  - There is also a simpler standard (part of XQuery, XSL, and XPointer)
    called XPath.

  - All languages are based on navigating through the structure
    of the XML document.


Ex: Find the titles of books costing < $60 where Ullman is an author
    (based on the first XML data)


In Lorel:

SELECT b.TITLE
FROM   BOOKSTORE.BOOK b
WHERE  b.@PRICE < $60 AND b.AUTHORS.AUTHOR.LASTNAME = "Ullman"


In XPath:

BOOKSTORE/BOOK[@PRICE<60, AUTHORS/AUTHOR/LASTNAME="Ullman"]/TITLE


In XQuery:

FOR $b IN BOOKSTORE/BOOK
WHERE $b/@PRICE < 60 AND $B/AUTHORS/AUTHOR/LASTNAME="Ullman"
RETURN
  <TITLE>$b/TITLE</TITLE>


XML query languages often include "wildcards" and regular expression
operators for cases when exact structure of data may be unknown.


EX: Find all titles anywhere in the bookstore containing "XML".

In Lorel:

SELECT t
FROM   BOOKSTORE.#.TITLE t
WHERE  t LIKE "%XML%"


Million (billion actually) dollar question
==========================================

* Will people store their data in XML, or only use XML as a transport
  format for data stored in conventional database systems?