Teaching XML databases with eXist

1
Teaching XML databases with eXist

• Chris Wallace
• Senior Lecturer
• School of Information Systems
• University of the West of England, Bristol, UK
• TLAD 2006
• 18th July 2006

2
Outline

• The teaching context
• The rationale for XML Databases
• eXist as a suitable Native XML database
• Conceptual data modelling
• Conceptual language modelling
• Issues

3
The teaching Context

• Multimedia Computing Programme
• First year Java, HTML, web applications with Perl
• Level 2 module Data, Schemas and Applications
• Broad coverage of data structures
• Tables
• Tagged structures
• Trees
• Sequences
• Web application development
• Emphasis on transferable knowledge

4
The workplan

• Semester 1
• Use ER models to capture understanding of Flickr,
  an online photo album
• Groups of 3 Use PHP and MySQL to implement a
  baby Flickr
• Semester
• Use XML to structure data (Whisky data)
• Use XQuery and eXist to develop individual
  XML-based application (jh)

5
The Rationale for XML database

• Case study the CEMS FOLD
• A faculty web site providing staff and students
  with information about
• Modules
• Programmes
• Staff
• Organisational structure

Word documents
Spread sheets
SQL DB
JPEGs
LDAP
6
Why not an RDBMS?

• Documents
• Unit of editing, ownership, communication
• Complex composite structures
• Ordering and other complex relationships
• Linked by shared values
• Spreadsheets
• Familiar data organisation tool
• Process descriptions
• eBXML

7
structured text
composition
Complex expressions
associations
ordering
subtypes
8
eXist Native XML Database

• Open source Java
• European team of developers led by Wolfgang
  Meier(Darmstadt)
• Documents (files) are organised in collections
  (folders) in a file store
• XML Documents stored in an efficient, B tree
  structure with indexes
• Non-XML resources (XQuery, CSS, JPEG ..), etc can
  be stored as binary
• Deployable in different ways
• Embedded in a Java application
• With embedded Jetty HTTPserver
• Multiple Interfaces
• REST to Java servlet
• SOAP
• XMLRPC
• Support for XQuery
• Function library extensions
• Language extensions (free-text matching and
  update)
• Schema validation supported
• Database is not schema-aware no type
  information available

9
NDX/RDBMS similarities and differences

• Similarities
• Associations implemented via common values
• Differences
• NDX can store documents without a prior schema
  agile modelling
• Composition implemented via containment
• Sequence directly supported
• XQuery fulfils the role of SQL (PL/SQL,PHP,JSP)

10
SQL and XQuery

• SELECT FROM emp WHERE deptno 10
• //Empdeptno 10
• SELECT empno, location
• FROM dept, emp
• WHERE dept.DEPTNO emp.DEPTNO
• for e in //Emp
• let d //Deptdeptno e/deptno
• return (e/empno,d/loc)
• SELECT e.ename,m.ename
• FROM emp AS e, emp AS m
• WHERE e.mgr m.empno
• AND e.deptno 20
• for e in //Empdeptno 20
• let m //Empempno e/mgr
• return ltempgte/ename ,m/enamelt/empgt

11
Simple Table Example

• Data on Excel spreadsheet
• Convert to XML (Excel add-in)
• Save as XML data
• Load into eXist database
• Run trip.xql to find and generate XQuery to
  generate .kml file
• View with GoogleEarth

12
Stored Query

• lthtmlgt
• ltheadgt
• lttitlegtBSA trip reportslt/titlegt
• lt/headgt
• ltbodygt
• lth1gtBSA trip reportslt/h1gt
• lttable border"1"gt
• lttrgtltthgtStart Datelt/thgtltthgtTriplt/thgtlt/trgt
• for trip in collection("/db/chriswallace/voyage")
  //trip
• return
• lttrgt
• lttdgtstring(trip/startDate)lt/tdgt
• lttdgtstring(trip/title)lt/tdgt
• lttdgtlta href"trip2kml.xql?idtrip/id"gtGoog
  le Earthlt/agtlt/tdgt
• lttdgtlta href"showtrip.xql?idtrip/id"gtShow
  triplt/agtlt/tdgt
• lt/trgt
• lt/tablegt

13
Conceptual Data Models

• QSEE allows conceptual data models to be created,
  and then SQL DDL generated with
• Foreign key generation
• Composite primary keys
• Many-many resolution
• Creates the necessary separation between the
  conceptual model and its realisation in a given
  logical data model
• The student to develop a sense of the utility of
  ER models across a wide range of structures

14
Conceptual Language Model

• Web development calls for multiple languages
  Java, JavaScript, Perl, PHP, XQuery, XSLT, SQL
• This provides the opportunity and need to try to
  dig down to the common conceptual models e.g.
• Formal/informal language mixture
• Structured programming
• Modularity

15
Issues

• Controlled exposure to language
• Complexity of XQuery versus SQL
• Schemas hindrance or lifeline?
• Tools for conceptual modelling
• Weaker students

16
Takeaways

• XQuery and the eXist Native XML database are the
  most fun Ive had since Smalltalk
• actually more because it delivers useful systems
• Take the code to the data, not the data to the
  code
• PHP/ Java
• loading and unloading XML ltgt internal structures
  (SAX, DOM)
• XQuery/XSLT
• Direct operations on the persistent XML data
• Content is king
• A successful Information system depends on
• Data quality model fidelity more than model
  integrity
• Handling multiple, sometimes inconsistent, views
  of the world