DASWIS 2001 NF-SS: A Normal Form for Semistructured Schemata Xiaoying Wu, Tok Wang Ling, Sin Yeung Lee, Mong Li Lee National University of Singapore Gillian Dobbie University of Auckland, New Zealand 1 DASWIS 2001 Outline 1. Motivations 2. Semistructured schema and its data tree 3. Integrity constraints for semistructured data 4. NF-SS: Normal Form for Semistructured Schemata 5. Designing of semistructured schema into NF-SS 6. Discussions of the designing approach 7. Comparison with related proposal 8. Summary 2 DASWIS 2001 1. Motivation: Example 1 <!ELEMENT department (course+) <!ATTLIST department name ID #REQUIRED> <!ELEMENT course (students*)> <!ATTLIST course cid ID #REQUIRED title CDATA #implied> <!ELEMENT student (grade?)> <!ATTLIST student sid ID #REQUIRED name CDATA #REQUIRED age CDATA #IMPLIED> <!ELEMENT grade (#PCDATA)> department name + course cid title * student ? sid age name grade 3 DASWIS 2001 1. Motivation (cont.) Redundancy: name and age of a student Updating Anomaly: – – – Insertion Rewriting Deletion department course name: CS course student sid: s01 name: age: 21 Jack title: database design cid: cid: cs5220 cs4221 student sid: s02 name: grade Tom sid: s01 title: data Mining student name: age: 21 Jack “A” 4 DASWIS 2001 1. Motivation:Example 2 <!ELEMENT teacher (ClassRoom*)> <!ATTLIST teacher tid ID #REQUIRED> teacher name CDATA #REQUIRED> <!ELEMENT ClassRoom (subject*)> name * tid <!ATTLIST ClassRoom room# ID #REQUIRED> ClassRoom <!ELEMENT subject (time)> * <!ATTLIST subject room# subject cid ID #REQUIRED> <!ELEMENT time EMPTY> * cid <!ATTLIST day CDATA #REQUIRED time hour CDATA #REQUIRED> day hour Path anomaly: –The schema doesn’t reflect the integrity constraints: tid,day,hourcid,room# 5 DASWIS 2001 2. Semistructured Schema and Data tree A semistructured schema is defined to be D = (E, A, B, P, R, r) •E is a finite set of object types in D. E: Object r: root type type Object •A is a finite set of attributes, disjoint from E. A: attribute •B is a set of basic domain type like string, integer, s Boolean etc. •P is a function from E to object type definition with symbol in {*, +, ? ,1} called multiplicity e.g: P (course) = student* •R is a function from E to the power set of A e.g.: R(student) = {sid, name, age } • r E and is called the object type of the root. department name + course multiplicity cid title * student ? sid age name grade e.g.: r = department 6 DASWIS 2001 2. Semistructured Schema and Data tree (Cont.) A data tree T with respect to a semistructured schema D = (E, A, B, P, R, r) is defined to be a tree T=(V, lab, obj, att, val, root), showing a database instance. department course name: CS course student sid: s01 name: age: Jack 21 title: database design cid: cid: cs5220 cs4221 student sid: s02 name: grade Tom sid: s01 title: data Mining student name: age: Jack 21 “A” 7 DASWIS 2001 2. Semistructured Schema and Data tree (Cont.) •The path of a node n in semistructured schema D is denoted as pathD(n). e.g.: PathD for student is /department / course / student •The path of a node v in data tree T is denoted as PathT(v) e.g.: PathT for student “s02” is /department / course/ student •The target set of node n in T, T[n], is {v: vV, nEA PathT(v)= PathD(n)}. e.g.: the target set T[student] includes nodes of students with sid “s02” etc. department department name + course name:CS course cid:cs4221 cid title * student title:database design student student ? sid age name grade sid: name: s02 Tom grade “A” 8 DASWIS 2001 2. Semistructured Schema and Data tree (Cont.) Two nodes from two data tree w.r.t schema D satisfy value equality iff – – they are attributes nodes with the same tag and the same value; or they are object nodes having the same tag and their children are pairwise value equal Two data trees T1 and T2 w.r.t schema D = (E, A, B, P, R, r), X E A. T1 and T2 agree on X, denoted as iff the following condition is hold: t1T1[X],t2T2[X], such that (t1=vt2) department course name: CS course cid: student cs4221 sid: s01 name: age: 21 Jack sid: s02 title: database design student name: Tom grade “A” cid: cs5220 sid: s01 title: data Mining student name: Jack age: 21 9 DASWIS 2001 3. Integrity Constraints for Semistructured Data Extended Functional Dependency(EFD) Let D = (E, A, B, P, R, r) be a semistructured schema, let X EA and Y EA. Y is extended functionally dependent on X, is denoted as XY. Let S denotes a set of data trees that are images of D, S satisfies XY, iff for any data trees T1, T2 in S, if they agree on every component in X, then they will agree on Y.that is, T1, T2 S((xX, T1=xT2) such that T1=yT2). Inference rule for EFD E1:(reflexivity) If YX, then XY, for any X, Y EA E2:(augmentation) if XY then XZYZ, for any X, Y, Z EA E3:(transitivity) If XY, YZ then XZ, for any X, Y, Z EA 10 DASWIS 2001 3. Integrity Constraints for Semistructured Data (Cont.) Notation:O1[@X1], …, Oi[@Xi],…,On-1[@Xn-1]On[@Xn] EFD XY is partial EFD: If there exists an X’X such that X’Y. Otherwise, is full EFD. e.g.: (1) course[@cid],student[@sid]student[@name] is partial EFD (2) student[@sid]student[@name] its full EFD XY is said to be coherent iff /X/Y is a path in D; otherwise it is called an incoherent EFD. e.g.:teacher[@tid], time [@day, @hour]subject[@cid] is an incoherent EFD, since /teacher / time /subject is not a path in schema. teacher tid room# cid name * ClassRoom * subject * time 11 day hour DASWIS 2001 3. Integrity Constraints for Semistructured Data (Cont.) If there exists ZEA, such that XY and YZ and Y is transitively extended functionally dependent on X via Z. e.g.: age is transitively dependent on course via student since X, then Z (1) course[@cid]student[@sid] (2) student[@sid]student[@age] and (3)student[@sid] course[@cid] department name + course cid title * student ? sid age name grade 12 DASWIS 2001 3. Integrity Constraints for Semistructured Data (Cont.) Theorem Let D = (E, A, B, P, R, r) be a semistructured schema, X, Y, Z E A. If Z is transitively dependent on X via Y, then there exists a data tree of D where a rewriting anomaly occurs upon updating the values of Z. department course name: CS course student sid: s01 name: age: 21 Jack title: database design cid: cid: cs5220 cs4221 student sid: s02 name: grade Tom “A” sid: s01 title: data Mining student name: age: 21 Jack 13 DASWIS 2001 3. Integrity Constraints for Semistructured Data (Cont.) : Based on EFD semantics Notation: Ko = O1[@X1]/…/Oi[@Xi]/…/On[@Xn]/O[@X] for key of an object type O in semistructured schema D. /O1/…/O is a path in D If n equals one, then Ko is called an absolute key. Otherwise it is called a relative key. Key Constraints Example book •Kbook= book[@isbn]. Kbook is an absolute key •Kchapter =book[@isbn]/chapter[@number]. relative key isbn Kchapter is a •Ksection= book[@isbn]/chapter[@number]/section[@number]. Ksection is a relative key + chapter number + section number 14 DASWIS 2001 3. Integrity Constraints for Semistructured Data (Cont.) Let D be a semistructured schema and O be its root object type. The set of basic dependencies of D, denoted as BD(D), is defined as follows: Let X, Y be children of O, non-trivial extended functional dependencies of the form XY where X is a key of O or Y is part of a key of O, are in BD(D). Let O1 be a sub-object type of O and D1 be a schema tree that is rooted at O1 and add KO as attribute(s) of O1, then BD(D1) BD(D). No other non-trivial dependencies that is not generated from above is in BD(D) 15 DASWIS 2001 4. NF-SS Let D be a semistructured schema and O be its root object type. D is in Normal Form for Semistructured Schemata (NF-SS), iff 1. O has at least one key. 2. For any non-trivial EFD of the form XY satisfied by O, where X and Y are attributes of O, then either X is a key or Y is part of the key of O 3. For any sub-object type O1 of O (a) If adding KO to O1 as its components with other remains, a schema tree rooted at O1 will be in NF-SS. (b) KO KO1= or KO KO1, where KO and KO1 are O and O1’s key respectively. (c) O1 is not transitively dependent on KO 4. Any non-trivial EFD in D can be derived from BD(D) by using the inference rules for EFDs. 16 DASWIS 2001 5. Designing Semistructured Schema into NF-SS We adopt restructuring approach for the designing. We propose four heuristic restructuring rules – Decomposition object types. – Creation new object types. – Regrouping components of an object type. Objective – Remove transitive or partial EFD and incoherent EFD from the given dependency and key constraints. 17 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Rule 1. (Remove Transitive Dependency by Decomposition) Given an object type O in a semistructured schema D, if there is some non-prime component(s) Y of O that is transitively dependent on some key of O, i.e., KO X, X Y and X KO , and X KO =. Then, restructuring the schema as follows. 1. Duplicate X to form a new node(s) Z. 2. Move Y and all the descendants of Y and their corresponding edges under Z. 3. Make X as foreign key of O, and add a reference edge from the original node X to Z. 18 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Example 5.1: schema D satisfies the following EFDs (1)department[@name]course[@cid] (2) course[@cid]department (3)course[@cid]course[@title] (4)course[@cid]student[@sid (5)course[@cid],student[@sid]grade (6)student[@sid]student[@name, @age] name department department + course + course cid title * student ? sid age name grade name student2 cid title * sid age name student1 ? sid grade 19 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Rule 2. Remove Path Anomaly by Path Splitting Given a semistructured schema D. Suppose there exists an incoherent EFD: O1[@X1],…,On[@Xn] Y, Y is either an object type or an attribute, and there exists a path P that contains {O1,…,On,Y}. Path P can be split into two sub-paths P1 and P2,where P1 only contains {O1,…,On } and Y, while P2 contains {O1,…,On} and (P-Y). 20 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Example 5.2:schema D satisfies following EFDs (1) teacher[@tid],timeClassRoom (2)teacher[@tid], timesubject teacher tid room# cid name * ClassRoom * subject * time day hour teacher tid * time name day hour ClassRoom room# subject cid 21 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Rule 3. Removing Partial Dependency by Creating New Object type Given an object type O in a semistructured schema, let X be a set of prime attributes of O, and Y be the set of O’s attributes. Let O1 be a sub-object type of O. If (KO -X) O1 and no proper superset of X satisfy this property, then restructure the schema as follows: 1. (KO Y –X) becomes the only attribute(s) of O while O1 remains to be its sub-object type. 2.Create a new object type O2 that is a direct component of O. 3.Move rest of the components of O and all their descendants and corresponding edges under O2. 22 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Example 5.3: schema D shown in Figure (a). the following EFDs {O[@A,@B]D, O[@A,@B]O2, O[@A] O1, O[@A] E } and the key of O is {A,B}. O A B * O1 C O' Rule 3 D E O[@K, @B]O2 O2 F (a)Un-normalized schema as the partial EFD O[@A,@B} O1 O'' Rule 2 O’[@A]E A O1 C * A * O3 B D E E * * O3 O1 * O2 F (b)Un-normalized schema as the incoherent EFD O’[@A]E C B D * O2 F (c)Normalized schema 23 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Rule 4. (Restructuring To Satisfy Condition 3(b) of NF-SS Definition) Given an object type O in a semistructured schema D, X be a set of O’s attributes and single-valued atomic sub-object types, O1 be a complex sub-object type of O. O1 has relative key KO1 , but KO KO1 and KO1 KO .Let Y be KO KO1 X, and Y . D is restructured as follows: 1. O1 remains to be a sub-object type of O. 2. Make Y as components of O. 3.Create a new object type O2 to be a child of O and the rest components of O (excluding Y) become children of O2. 24 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Example 5.4: schema D in Figure (a) satisfies the EFD (1) O[@K, @A] O1 (2) O[@K, @B]O2 and the key of O is {K, A, B}. O' Rule 3 O O[@K, @B]O2 K A * O1 * K A * O2 C C D * O3 O1 B D B O'' Rule 4 o o 3 K * O4 * O2 * O1 A * O3 B * O2 E F (a)Un-normalized schema as O1 and O2 partially dependent on {K,A,B} E F (b)Un-normalized schema as KO=O’[@K,@A] and KO3=O’[@K]/O3[@B] such that KO KO3 C D E F (c)Normalized schema 25 DASWIS 2001 5. Designing Semistructured Schema into NF-SS(cont.) Algorithm 1: Restructuring Algorithm Input: A set S that contains semistructured schemas, and a set of EFDs for S. Output: A set of semistructured schemas that in NF-SS. Begin 1. for each semistructured schema D in S do if D is not in NF-SS then repeat until no further change: (1) if there exists transitive EFD: KO X, X Y and X KO for an object type O in D, Case X KO =: apply Rule 1 to remove the transitive EFD. Case X KO : apply Rule 3 to remove the transitive EFD. Case X KO : apply Rule 4 to remove the transitive EFD. (2) if there exists incoherent EFD then apply Rule 2 to remove it. 2. output S. End 26 DASWIS 2001 6. Discussion of Restructuring Approach for Designing Is the restructuring rules complete? No. – covering is not guaranteed – dependency preservation is not guaranteed Does it give unique solution? No. – depending on the order in which the dependencies are examined Designing task can be made easier if more semantics available. – In [5], We have proposed another approach for designing semistructured databases using ORA-SS, a semantic rich model . Nevertheless, it does give practical heuristics and provides insights into the normalization task for semistructured databases. 27 DASWIS 2001 7. Comparison with Related Proposal The first attempt to define normal form for semistructured data ([ER’99] S.Y.Lee, M.L.Lee, T.W.Ling, and L.A.Kalinichenko.) [3] – Defines a schema called S3-Graph, which makes no distinction between element node and attribute node and no cardinality specification. – Proposes S3-NF, but missing key constraints, an essential part of database design. – The decomposition method may not be able to remove some other kinds of anomalies, like partial dependency and path anomaly that may exist in a schema. The most recent proposal: XNF (XML Normal Form) ([ER 2001] D.W.Embley and W.Y.Mok. ) [2] – It mainly provides algorithms to translate a schema, represented in a conceptual model called CM hypergraphs, to a scheme-tree forest in XNF. – Like S3-Graph, scheme tree doesn't lend itself to XML definition. – XNF isn’t formulated with the concept of key. – The algorithms given suffers from efficiency. – A large set of results is expected. 28 DASWIS 2001 8. Summary A normal for semistructured schemata – It is incorporated with integrity constraints. – It guarantees no redundancy and hence no undesirable updating anomalies for the conforming semistructured databases. – It gives more reasonable representations of real world semantics Restructuring Approach for designing semistructured databases – a set of heuristic restructuring rules is proposed. – an algorithm for iteratively restructuring a schema into NF-SS is developed. – It provides insights into the normalization task for semistructured databases. 29 DASWIS 2001 References 1. J. Clark and S. DeRose. XML Path Language (XPath). W3C Working Darft, November 1999. http://www.w3.org/TR/xpath. 2.D.W.Embley and W.Y.Mok. Developing XML Documents with Guaranteed “Good” Properties. Proceedings of the 20th International Conference on Conceptual Modeling (ER), 2001. 3. S. Y. Lee, M. L. Lee, T. W. Ling and L. A.. Kalinichenko. Designing Good Semi-structured Databases. Proceedings of the 18th International Conference on Conceptual Modeling (ER), 1999. 4. T. W. Ling and L. L. Yan. NF-NR: A Practical Normal Form for Nested Relations. Journal of Systems Integration. Vol4, 1994, pp309-340 5. Xiaoying Wu, Tok Wang Ling, Mong Li Lee, Gillian Dobbie. Designing Semistructured Databases Using the ORA-SS Model, accepted for publication in Proceedings of the 2nd International Conference on Web Information Systems Engineering (WISE) , IEEE Computer Society, Kyoto, Japan, December 2001. 30 DASWIS 2001 Q&A 31