1. RRXS Redundancy reducing XML storage in relations O. MERT ERKUŞ 2002701054 A. ONUR DOĞUÇ 2002701069

2. INTRODUCTION FUNCTIONAL DEPENDENCIES CONSTRAINT PRESERVING RELATIONAL STORAGE EXPERIMENTAL EVALUATION CONCLUSION PRESENTATION OUTLINE

3. INTRODUCTION

4. PROBLEM  Current techniques for storing XML using relational technology, consider the structure of an XML document but ignore its semantics. However, when the semantics of a document is considered redundancy may be reduced! However, when the semantics of a document is considered redundancy may be reduced!

5. INTRODUCTION RELATIONAL DATABASES REVIEW STRUCTURAL APPROACH DTD – SCHEMA GRAPHS STRUCTURAL + SEMANTIC APPROACH KEYS – FOREIGN KEYS FUNCTIONAL DEPENDENCIES XML SCHEMA

6. INTRODUCTION PROBLEM DEFINITION Providing a mapping from XML to a relational database taking structural as well as a broad class of semantic constraints into account. Providing a mapping from XML to a relational database taking structural as well as a broad class of semantic constraints into account.

7. INTRODUCTION EXAMPLE - XML TREE

8. INTRODUCTION EXAMPLE - CONSTRAINTS

9. INTRODUCTION EXAMPLE - COMMENTS  1. & 3. constraints are STRUCTURAL. 2. & 4. constraints are KEYS 2. & 4. constraints are KEYS 5.constraint is FUNCTIONAL DEPENDENCY 5.constraint is FUNCTIONAL DEPENDENCY  None of the relational storage strategies designed to date would produce this design.

10. INTRODUCTION OUTLINE OF THE WORK 1.A new constraint definition, XFDs, that can capture structural and key constraints, as well as the functional dependencies 1.A new constraint definition, XFDs, that can capture structural and key constraints, as well as the functional dependencies 2.A set of rewriting rules for XFDs 2.A set of rewriting rules for XFDs 3.A polynomial time algorithm to reduce the input set of XFDs 3.A polynomial time algorithm to reduce the input set of XFDs 4.A constraint-preserving mapping into relational storage that reduces redundancy 5.Experimental evaluation which shows the effectiveness of RRXS

11. FUNCTIONAL DEPENDENCIES

12. DEFINITION  Functional dependencies for XML (XFDs) are used to describe the property that the values of some attributes of a tuple uniquely determine the values of other attributes of the tuple.

13. FUNCTIONAL DEPENDENCIES EXAMPLE – XFD’s from constraints Variable Bindings $x in //vendor $y in //book $z in $x/book

14. FUNCTIONAL DEPENDENCIES DEFINITIONS - SUITE An “attribute” for XML, called a P-attribute, is defined by a path expression $v=Q that occurs in some functional dependency.An “attribute” for XML, called a P-attribute, is defined by a path expression $v=Q that occurs in some functional dependency. The set of P-attributes in an XFD group together values to form a ‘tuple’ for an XML instance, named an X-tuple.The set of P-attributes in an XFD group together values to form a ‘tuple’ for an XML instance, named an X-tuple. A functional dependency is defined on the P-attributes of an X-tuple, and intuitively must hold on the set of all X-tuples formed by valid variable bindings. A functional dependency is defined on the P-attributes of an X-tuple, and intuitively must hold on the set of all X-tuples formed by valid variable bindings.

15. FUNCTIONAL DEPENDENCIES TYPES OF XFD’S  Structural XFD’s : Structural XFDs are used to capture the tree structure of an XML document and certain types of schema information. C1, C3  Semantic XFD’s: Semantic constraints are used to capture deeper knowledge of the data. C2, C4, C5

16. FUNCTIONAL DEPENDENCIES REDUCING XFD’S THE TASK: FINDING A SET OF RULES, WHICH CAN PROVE THE SOUNDNESS & COMPLETENESS OF THE XFD INFERENCE THE TASK: FINDING A SET OF RULES, WHICH CAN PROVE THE SOUNDNESS & COMPLETENESS OF THE XFD INFERENCE

17. FUNCTIONAL DEPENDENCIES REWRITE RULES 1.Armstrong Axioms ReflexivityAugmentationTransitivity 2.Containment To use path expressions instead of simple attributes. Considers the relationship between path expressions.

18. FUNCTIONAL DEPENDENCIES REWRITE RULES 3.Singleton path 4.Variable-move 5.Variable Introduction and Elimination Exploits structural constraints imposed by the definition of XFDs. Move variable bindinds in relations Insert new variables and eliminate redundant ones

19. FUNCTIONAL DEPENDENCIES XFD INFERENCE INFER: A polynomial time algorithm which “Given an XFD Ø : X  Y and a set of XFD’S F, determines wheter or not Ø can be inferred from F using L (Rewrite Rules).” It detects which XFD’s can be eliminated or simplified, from the initial set of XFD’s and derives G (Redundancy reduced set of XFD’s)

20. CONSTRAINT PRESERVING RELATIONAL STORAGE

21. RRXS: SCHEMA MAPPING The XML-to-Relational mapping method has following input and outputs : Input: A set ofXFDs F, and an optional DTD D. Output: A target relational schema R with a set of keys K, and a redundancy reducing, constraint preserving transformation M.

22. CONSTRAINT PRESERVING RELATIONAL STORAGE RRXS: SCHEMA MAPPING REDUNDANCY REDUCING: It means that: redundancy which can be detected by F using L is eliminated in R. CONSTRAINT-PRESERVING : It means that : for any XML tree T, F hold on T if and only if K hold on M(T).

23. CONSTRAINT PRESERVING RELATIONAL STORAGE ALGORITHM RRXS

24. CONSTRAINT PRESERVING RELATIONAL STORAGE ALGORITHM RRXS EQUIVALENCE: An algorithm to recognize equivalent XFDs and equivalent elements, then group them in equivalence classes and output G. REDUCE: An algorithm similar to ‘infer’ removing redundant XFDs SHRINK: An algorithm that removes unnecessary elements, producing the set of XFDs

25. CONSTRAINT PRESERVING RELATIONAL STORAGE INSTANCE MAPPING The instance mapping takes an XML tree T which conforms to DTD D and satisfies the XFDs F as well as the schema mapping output M, and generates a relational instance M(T) which conforms to schema R.

26. EXPERIMENTAL EVALUATION

28. RESULTS 1.SOME NODE IDS GENERATED BY HYBRID INLINING ARE REMOVED. (ID) 2.USER DEFINED XFD’s ARE CORRECTLY USED TO ELIMINATE REDUNDANCIES 3.THE STRATEGY WORKS CORRECTLY FOR RECURSIVE DATA

29. EXPERIMENTAL EVALUATION RESULTS

30. CONCLUSION SUMMARY AND FUTURE WORKS