Articles & Content

Archives

Relational Data Modeling With pureXML (E07)

Topic: Cross Platform DB2 for z/OS & LUW

Subtopic: 2010



DATE: 2010-11-9 (15:45 - 16:45)
SPEAKERS: Bayard Tysor (Bayard Lee Tysor, Inc.)

Incorporating XML, a hierarchical data structure, into a relational data model sounds like an oxymoron, however now that IBM has incorporated pureXML into DB2, it is necessary to consider what should be stored relationally, and what should be included in an XML column, or possibly both. After a quick overview of pureXML and normalization, this presentation will explore data architect's choices in storing XML data. This includes 1) not in DB2, 2) in character string or CLOB columns, 3) shredding into fully relational, 4) fully storing in an XML column, or 5) some hybrid design. The advantages and disadvantages for each of these design decisions will be stressed. These choices should be made based on the need for optimal quick responses to schema changes, vs. performance (both CPU and elapsed time), vs. programmer coding productivity. There will be examples to reinforce these concepts.pureXML overviewNormalization OverviewKeeping in XML onlyFully Shredding into RelationalHybrid Approach

EXP. LEVEL: Intermediate,Advanced

OBJECTIVES:

pureXML overview: An objective of this presentation will be to quickly review how DB2 parses and stores XML. This will include SYSXMLSTRINGS, separate ‚ úLOB‚ ? type tablespace, and special indexes.

Normalization Overview: An objective of this presentation will be to review relational design normalization. This includes definition and examples of the ridged first, second and third normal forms, as compared to flexible and hierarchical XML.

Keeping in XML only: An objective of this presentation is to show the advantages and disadvantages of storing an XML document intact in an XML column, and how this leads to maximum flexibility and minimum performance.

Fully Shredding into Relational: An objective of this presentation is to show the advantages and disadvantages of fully shredding XML data in to fully relational columns and how this leads to maximum performance and minimum flexibility.

Hybrid Approach: An objective of this presentation is to explore the design decisions involved in determining which portions of an XML document to shred into relational columns for performance, and what to keep in XML column(s) for flexibility.



Click Here to Download

NOTE: These are only open to members of IDUG. If you are not a member, please CLICK HERE for more information.