Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.

3, 2012

www.iiste.org

Gn-Dtd: Innovative Way for Normalizing XML Document

*[email protected] , [email protected]

AbsractAs XML becomes widely used, dealing with redundancies in XML data has become an increasingly important issue. Redundantly stored information can lead not just to a higher data storage cost, but also to increased costs for data transfer and data manipulation, such data redundancies can lead to potential update anomalies. One way to avoid data redundancies is to employ good schema design based on known functional dependencies. This paper presents a graphical approach to model XML documents based on a Data Type Documentation called Graphical but precise. Using relationship, elements Notations-Data Type Documentation in a and (GN-DTD). simple way GN-DTD allows us to capture syntax and semantic of XML documents attributes, between hierarchical structure, cardinality,

various notations, the important features of XML documents such as elements, sequence disjunction

or attribute are visualize clearly at the schema level.

Keywords- XML Model, GN-DTD design, Normalization XML schema, Transformation Rules

1.INTRODUCTION With the wide exploitation of the web and the accessibility of a huge amount of electronic data, XML (extensible Mark-up Language) has been used as a standard means of information representation and exchange over the web. Additionally, XML is currently used for many different types of applications which can be classified into two main categories [5,6]. The first application is called document centric XML and the other is called data centric XML. The document centric XML is used as a mark-up language for semi-structured text documents with mixed-content elements and comments. The data centric XML consists of regular structure data for automated processing and there are little or no element with mixed content, comments, and processing instruction. The current XML data models however do not pay sufficient attention to the Problem of representing the structure of XML documents. We believe, in order to present more sophisticated forms of XML documents structure, the schema such as DTD or XML schema must taken into account since it is used to define and validate XML documents structure. In our work, we consider DTD, as it has been widely well accepted and expressive enough for a large variety

29

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
of applications.

www.iiste.org

Furthermore DTD is an early standard for XML, and many legacy XML documents

structures are defined by DTDs. In this paper, we proposed a graphical notation of DTD called GN-DTD to overcome the above limitations. The GN- DTD helps to arrange the content of XML documents in order to give a better understanding of DTD structures, to improve an XML design and normalization process as well. GN-DTD has richer syntax and structure which incorporate of attribute identity, simple data type, complex data type and relationship types between the elements. Furthermore, the semantic constraints that are important in XML documents are defined clearly and precisely to express the semantic expressiveness. 2. RELATED WORK Major current XML data models use directed edge labelled graphs to represent XML documents and their Schemas .These models consist of nodes and directed edges which respectively represent XML element in the document and relationship among the element. These existing XML model can be categorised into:XML model to represent instance of XML document,XML model represent XML schema

are DOM(document object model),OEM(object exchange model)[7],S3-GRAPH[2] and many more.

models is on modelling the nested structure of semi structured data but not modelling the constraint that hold in the data. In constrast, data model such as S3-Graph, CM Hyper graph, EER, XML Trees and ORA-SS have been defined specifically for data management. Amongst these models, the notation of ORA-SS, semantic network model and EER notations are best to be adopted and applied in GN-DTD. 3.XML MODEL DESIGN Consider the DTD in Fig. 1 The first line of DTD in Fig. 1 shows that department is the root of

the DTD. While second line shows that department consists of sub element course. The semantic relationship between department and course is indicated by the symbol *, represents that department can consists of zero or many course for each department. The third line of the DTD shows that each element course has sub element title and element taken_by. Symbol , between them indicated that they must occur in sequence. The fourth line indicates that element course has an attribute cno. The keyword #REQUIRED represents that the attribute cno must appear in every course while ID indicates that the value of cno is unique within XML document. The fifth line of the DTD shows that the keyword PCDATA to despite that element title has no sub element and it is a leaf element and has a string value. <!DOCTYPE department[ <!ELEMENT department(course*)> <!ELEMENT course(title,taken_by)>

30

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
<!ATTLIST course cno ID #REQUIRED> <!ELEMENT title (#PCDATA)> <!ELEMENT taken_by (student*)> <!ELEMENT student(firstname|lastname?,teacher)> <!ATTLIST student Sno ID #REQUIRED <!ELEMENT title (#PCDATA)> <!ELEMENT taken_by (student*)> <!ELEMENTstudent(firstname|lastname?,teacher)> <!ATTLIST student Sno ID #REQUIRED

www.iiste.org

<!ELEMENT firstname(#PCDATA) > <!ELEMENT lastname(#PCDATA) > <!ELEMENT teacher (tname)> <!ATTLIST teacher tno ID #REQUIRED <!ELEMENT tname (#PCDATA) Fig1:DTD STRUCTURE DESIGN ITS related XML document confirms to dtd is as follows <!DOCTYPE courses [ <courses> <course> <course cno = csc101> < title > XML database </title> <taken_by> < student > <student sno = 112344> <firstname> zurinahni</firstname> <lastname> zainol </lastname> <teacher> <teacher tno = 123> <tname>Bing </tname> </teacher> </student> < student > <student sno = 112345> <firstname>Azli </firtname> <teacher> <teacher tno = 123> <tname> Bing </tname> </teacher> </student> <course> <course cno = csc102> < title > Database Design </title> <taken_by> < student > <student sno = 112344> <firstnme> zurinahni</firtname> <lastname>zainol </lastname> <teacher>

31

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
<teacher tno = 123> <tname> Botaci </tname> </teacher> </student> < student > <student sno = 112345> <firstnme>Azli </firstname> <teacher> <teacher tno = 123> <tname> Botaci </tname> </teacher> </student> </course> </courses> Fig2: XML document related to above DTD Any XML contain document that satisfies and conforms to this DTD is

www.iiste.org

likely

to

data redundancies which may lead to update anomalies. For example, as shown in Figure 2,

the lecturer named Bing who teaches the same course number (cno) csc101 is stored twice, which will lead to the updation anomalies. To avoid such problems, a set of rules should be provided when designing a DTD for XML documents. 4.TRANSFORMATION OF DTD INTO GN-DTD GN-DTD emphasizes the representation of semantic constraints between the complex elements, simple elements and attributes clearly. GN-DTD represents the structure and the semantic constraints of the XML document in a schema level. GN-DTD has following basic components: Aset of complex element node representing the element that have subelement A set of simple element nodes epresenting simple element that have no subelement A set of attributes nodes representing the attributes defines in ATTLIST. A semantic relationship between two nodes. A root node

Consider following DTD <!DOCTYPE department[ <!ELEMENT department(course*)> <!ELEMENT course(title, student*)> <!ATTLIST course cno ID #REQUIRED> <!ELEMENT title (#PCDATA)> <!ELEMENT student(fname|lname?,lecturer)> <!ATTLIST student Sno ID #REQUIRED <!ELEMENT fname(#PCDATA)> <!ELEMENT lname(#PCDATA)> <!ELEMENT lecturer(tname)> <!ATTLIST lecturer tno ID #REQUIRED> <!ELEMENT tname (#PCDATA)>

32

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
]> Fig 3: DTD Formation Following is the list of some notations used to representGN-DTD

www.iiste.org

5. Constrant Between Set Of Relationship

5.1 Sequence Between Set Of Child Element Nodes

Normally each complex element node consist a single attribute node or multi attribute node. We emphasize in our notation those node must be located first in the sequence before include other simple or complex elements node. To illustrate this, we draw a directed curved up arrow and labeled with {sequence} across all the set of relationship involved. Consider the following segment of DTD and its GN-DTD where attribute Sno is located at first position in the sequence of child elements. <!ELEMENT student (fname,lname,grade)>

33

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
<!ATTLIST student Sno ID #REQUIRED> <!ELEMENT fname(#PCDATA) > <!ELEMENT lname(#PCDATA) > <!ELEMENT grade(#PCDATA) >

www.iiste.org

Fig 4:Sequence of Attributes

5.2.Sequence Between The Set Of Sub Element We have a set of sub elements that are in an exclusive OR {XOR} relationship to represent notation |in DTD. For example, for the complex element node student, only one of its sub elements which are fname or lname, to be appeared as its sub elements in the XML document. To illustrate this, we draw a line and labeled with {XOR} across all the set of relationship involved. Follows is a real example of application . <! ELEMENT chapter (page| citation| table)* > which is equivalent with<! ELEMENT chapter (page*| citation*| table*) >.

Fig 5:Disjunction of several Simple Element

Following is the GN-DTD formation of DTD in fig 3

34

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012

www.iiste.org

fig6:GN-DTD formation TO Better understand ,consider the following DTD <!DOCTYPE school[ <!ELEMENT school (course*|subject*)> <!ELEMENT course(students*)> <!ATTLIST course cno ID #REQUIRED> <!ELEMENT subject(students*)> <!ATTLIST subject sno ID #REQUIRED> <!ELEMENT students (student*)> <! ELEMENT student ( tel?, address*,grade?)> <! ATTLIST student Sno ID #REQUIRED> Name CDATA #REQUIRED> <! ELEMENT tel (#PCDATA)> <! ELEMENT address (EMPTY)> <ATTLIST address Code (CDATA) #REQUIRED street (CDATA) #IMPLIED city(CDATA)#REQUIRED> <! ELEMENT grade (#PCDATA)>

35

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012

www.iiste.org

This is The main Diagramatical Representation of DTD on which we are going to apply the Normalization Rules to delete all the redundancies,anomalies which makes the XML as a bad XML document. 6. NORMALIZATION RULES FOR GN-DTD 6.1 First Normal Form GN-DTD(1XNF GN-DTD) The first normal form for GN-DTD is about finding unique identifier attributes for the complex elements set, and checking that no node (complex element, simple element or attribute) actually represents multiple values. To be in first normal form, each attribute, complex element or simple element is not NULL and has a single label. More importantly, the primary key (unique identifier) for the complex element must be defined. a)Only one value for each simple element node or attribute node of GN-DTD can be stored. If there is more than one value, we must add some new element nodes or attribute nodes to store them. b)The root element of a GN-DTD model should be located at level 0 and the cardinality of the root element node must be one. c) Each set of complex element node in the GN-DTD has at least one key attribute node. 1.6.2 Second normal form (2XNF GN_DTD) Some nodes need to be restructured. However they can then still be in a single GN-DTD. This is

possible in XML because XML supports hierarchies in a single document, while relational databases do

36

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012

www.iiste.org

not support hierarchies in a single row. This is different from the relational second normal form (2NF), which requires one-to-many relationships to be in separate tables. The GN-DTD is in second normal form if and only if: a) GN - DTD is in 1XNF. b) There is no nested binary inheritance relationship or ternary inheritance relationship under many-to-many or one -to-much inheritance relationships with the following condition:For each nested set of complex element<CE,l+1> of <CE,l>, and any key attribute (ATT) of <CE,l>, the key attribute and simple element of <CE,l+1> is not partial dependent on ATT of complex element<CE,l> 1.6.3Third normal form (3XNF GN_DTD) In the third normal form of the GN-DTD,making changes to one unique complex element node set would not affect the integrity of another complex element node sets.If needed,acomplex element node set would be divided into two separate complex element node set. GN- DTD is in third normal form if and only if: a) GN-DTD is in 2XNF. b) There exists no nested inheritance relationship type of n-ary many-to-one or many-to-many under a one-to-many inheritance relationship set in GN-DTD and the following conditions are satisfied: (i)For each nested set of complex elements<CEb,l+1> of set of complex element<CEa,l>, any key attribute and simple element of <CEb,l+1> is not transitively dependent on ATT of complex element<CEa,l> (ii) Any key attribute node of any complex element node located in a different level are disjoint (ATT<CE,l> ATT<CE,l+1> ATT<CE,n> =0) 1.6.4 Normal form GN-DTD(NF GN-DTD) GN- DTD is in Normal Form if and only if: a) b) GN-DTD is in 3NF. There are no global dependencies between attribute and simple element of complex element

nodes under nested one-to-many or many-to-many inheritance relationship. 7. TRANSFORMATION FROM GN-DTD TO DTD After removing all the types of redundancies GN-DTD can be transform back to DTD structures Following is the set of some transformation rules used to come back to the original DTD Step 1 Level 0, a root node node is represented [element type definition] > By <!DOCTYPE root relationship type Step 3 If there is no more than one node at level 1and nodes are hierarchical then generate <!ELEMENT root node name ( Ni) )> Where Ni is the list of sub elements/child nodes 3.1 Check the relationship set between parent Nodes and child nodes, name

Step 2 Level 1, identity the sub tree of GN-DT check the number of nodes, type of nodes and

37

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
3.1.1 If {XOR} means the relationship between node symbol |Else 3.1.2 If {sequence} means the relationship

www.iiste.org

is a disjunction and will be represented using

is sequence and will be represented using symbol ,

3.2 Check the semantic constraint between parent nodes and child nodes in each of relationship set and map to following operator: 3.2.1if [0..N] map to operator *, 3.2.2if [1..N] map to operator + 3.2.3if [0..1] map to operator ? Step 4 If the list of sub elements (Ni) is not empty, using depth first traversal, for each node in list sub element Ni 4.1 repeat step 3.1 and 3.2 4.2 generate < ! ELEMENT Ni (sub element Nj)> 4.3 for each complex element (Ni), find an attribute node and generate <! ATTLIST Ni attribute name attribute type> 4.4 For sub element Nj with Ni then generate 4.4.1If Nj is a simple element has part of link <!ELEMENT simple element name #PCDATA> (Repeat for all simple element nodes) 4.4.2 If Nj is a complex element node has inheritance link with Ni Repeat step 4 4.4.3 If Nj is a complex element node has part of link then generate <!ELEMENT Nj (EMPTY) > Step 5 Go to next sub tree GN-DTD and repeat step 4 1.7 CONCLUSION We have proposed a method for designing a good XML document in two steps: first, we building a conceptual model by means of GN-DTD at the schema level and second, using normalization theory where functional dependencies are refined among its simple elements and attributes. The GN-DTD can be further normalised either to 1XNF, 2XNF, 3XNF or XNF using the proposed normalization algorithm. In the proposed methodology, a GN-DTD is used as input and the normalization rules are applied during the normalization process. We also explain the process for transforming GN-DTD into DTD. 1.8 REFERENCES [1] Areanas M. And Libkin , L. A Normal Form For XML Document ACM Transaction on Database 232 of relational and XMLdata,Journal of computer And System Vol29(1),2004,pp. 195system sciences,2007 [3] Ling,T.W,A normal Form for Entity-Relationship diagram,proceeding 4th International Conference on E-R Approach,1985,pp,24-35

[2] Kolahi,S., Dependancy preserving normalization

38

Computer Engineering and Intelligent Systems ISSN 2222-1719 (Paper) ISSN 2222-2863 (Online) Vol 3, No.3, 2012
[4] Ling,T.W., Lee,M.L.and Dobbie,G.SemiStructured Database Design,Springer2005

www.iiste.org

[5] Vincet, m., Liu,J.,Mohania,M.,On the equivalence Between FDs in XML and FDs in relations Actal Informatica,2007,pp,230-24 [6] Wang,j.and Topor,R.,Removing XML data reduncies Using Functionality Equqlity Generating Dependencies 16th Australasian database Conference,2005,pp,65-74 [7] Biskup,J.,Achievement of relational Dataase Schem Design theory revisited,Semantic in Database,LNCS Vol 1066,Springer,1995,pp,14-44 [8] Zainol,z.and Wang ,B.,GN-DTD:Graphical notation forDescribing XMl Document ,2nd International Conference on Advances in Databases,Knowledge.And Data Application,IEEE,2010

Author Biography: 1] , Miss. Jagruti Wankhade B.E.(I.T.), M.E.(I.T.) (appearing) sipnas college of Engg and Tech,Amravati S.G.B .Amravati University,(MS),India 2] Prof. Vijay Gulhane B.E.(CMPS), M.E.(CMPS),PhD (pursuing) S.G.B .Amravati University,(MS),India Working as a (A.P.) in sipnas college of Engg and Tech,Amravati

39