JP5090481B2 - Data modeling method, apparatus and program - Google Patents

Description

  The present invention relates to a data modeling method, apparatus, and program, and more particularly to data for designing a specific XML data structure in designing a system that handles structured data such as a large number of XML data that is structured data. The present invention relates to a modeling method, apparatus, and program.

  When designing an information system that requires management of a certain amount of data or more, a database management system (DBMS) is used, and for this purpose, design work of a data structure managed by a system called data modeling is usually performed.

  For data modeling, the following three types of data modeling are designed sequentially for the DBMS used in the system.

(1) Conceptual data model:
An abstract data model that does not depend on DBMS and describes the semantic aspects of information.

(2) Logical data model:
A data model that depends on the model targeted by the DBMS. For example, in the case of a relational database (RDB), it becomes a relational model composed of tables, columns, etc., and is transformed from the conceptual data model. .

(3) Physical data model:
It determines the physical internal structure of the database, and is transformed from the logical data model in consideration of performance.

  Designing these database models is called conceptual design, logical design, and physical design, respectively.

  In general, as a data modeling technique, an entity relation diagram (ER diagram) (see, for example, Non-Patent Document 1) or a unified modeling language (UML) class diagram is used. Conventionally, a relational database (RDB) is used for the DBMS.

  And the support tool for the above technique is prepared, ER diagram can draw conceptual, logical and physical data models, and automatically outputs RDB schema (table definition, column definition DDL) from there, Some have such functions.

  The following is an explanation of data modeling for RDB, which is a conventional method, using ER diagrams.

<Conceptual design>
First, in conceptual design, entities (managed objects such as events / people / things), relationships (associations between entities), and attributes (data items of entities) are identified. As an example, a conceptual data model of a system for managing school students is shown as an ER diagram in FIG. Such an ER diagram can be drawn using a support tool.

  As a notation, a notation called IDEF1X is used. In FIG. 18, a group of information such as “class”, “student”, “club activity”, and “examination result” is an entity. In the figure, the line connecting the entities represents the relationship, and the black circle represents the “N” side where the entity relationship is 1: N (one-to-many). In general, the “1” side is called “parent (entity)”, and the “N” side is called “child (entity)”. Therefore, the relationship between “class” and “student” is 1: N, and the relationship between “student” and “test result” is also 1: N. The relationship between “club activities” and “students” with black circles on both sides represents M: N (many-to-many). What is written in the entity is an attribute, and the attribute (group) on the separator is called a “primary key”, which is an item that can uniquely determine information.

  Since the conceptual data model is required to be in a normal form that avoids data redundancy and inconsistencies, in the actual design work, entity identification, relationship identification, normalization (operation to normalize), etc. Repeat the creation and modification process according to the procedure.

  Further, FIG. 19 shows the same data model represented by a UML class diagram. Since it is possible to describe almost the same information as the ER diagram, data modeling is often performed using this information.

<Logical design>
Next, in the logical design, transformation from the conceptual data model is performed so as to satisfy the constraints of the relational model (table model). For example, in the relational model, the relationship of M: N cannot be expressed as it is, and it is necessary to create a new entity in the middle and create a relationship in 1: N.

  In RDB, an entity corresponds to a table, and an attribute corresponds to a column. Relationships are expressed by adding attributes called foreign keys. An example of an ER diagram obtained by transforming FIG. 18 for RDB is shown in FIG. An entity “club activity-student relation” is added between M: N related club activities and students. In addition, the attribute written as “(FK)” is a foreign key attribute. For example, a student's class code (FK) holds a relationship by holding the same value as the class code that is the primary key of the class. It means to express.

  Some support tools automatically (semi-) create such a logical data model from the conceptual data model in FIG. 18 (editing is necessary because newly added entity names are temporary names) Therefore, it is expressed as semi-automatic).

<Physical design>
Physical design is performed mainly for the purpose of improving performance. In general, a process with high cost in RDB is a combination process.

  For example, in the case of frequently performing the process of simultaneously acquiring “student” and “test results” as shown in FIG. 20, a join process between two tables of “student” and “test results” is required, and acquisition takes time. This is a concern. Therefore, in the physical design, the normal form is intentionally broken, and the “test results” are included in the “student” entity (= table) (table integration) to select a data model with improved performance. May do. An example is shown in FIG. In this case, there is also a restriction that only “three test results” can be retained.

  After editing the ER diagram of the physical data model from the logical data model using the support tool, the RDB schema can be automatically output.

<XML and XMLDB>
XML (Extensible Markup Language) is one of markup languages, and data described in XML is structured and meaningful in structure. FIG. 22 is a diagram showing XML data and its structure. Therefore, XML data is a tree structure model, and each section is called a node. In particular, the root node is the root element, the value (description content) is a text node, the one described in the tag is the attribute (to distinguish it from the attribute of the data model), the text node and XML Nodes other than attributes are called elements. The entire data is called an XML document.

  A database having a function of managing such XML data is XMLDB.

  Compared to the relational model of RDB, it is an advantage that it has a high expressive power. Therefore, it is a disadvantage that it is difficult to handle.

Chen, Peter P. (1976). "The entity-Relationship Model-Toward a Unified View of Data", ACM Transactions on Database Systems, Vol. 1, No.1, 1976, pp. 9-36.

  In conventional RDB data modeling, there is a case where the normal form is destroyed in order to improve performance in physical design. If the normal form is broken, there is a problem that maintainability such as ensuring data consistency is lowered. XMLDB is more expressive than RDB, but it is difficult to handle.

  The present invention has been made in view of the above points, and an object of the present invention is to provide a data modeling method, apparatus, and program capable of achieving both maintaining a normal form and improving performance in physical design.

  FIG. 1 is a principle configuration diagram of the present invention.

The present invention (Claim 1) is a data modeling apparatus for designing an XML data structure of a database (XMLDB) that is connected to a user terminal 300 and stores XML of a structured document.
A logical model design unit 110 that defines a logical XML corresponding to each entity of an actual relationship diagram (ER diagram) representing a logical data model, and stores the logical XML definition information in the logical XML definition information storage unit together with the information of the ER diagram as logical XML definition information;
An access pattern input unit 121 that acquires an access pattern of the input application and stores it in the access pattern storage unit 140;
Based on the logical XML definition information stored in the logical XML definition information storage unit 130 and the access pattern stored in the access pattern storage unit 140, a basic format for expressing a relationship using a foreign key when an individual update process is accessed for each entity. And an XML collection design unit 122 that stores the entity in the physical data model storage unit 150 as an integrated format that represents a plurality of entities in a unified tree structure when the plurality of entities are accessed together,
An entity is acquired from the physical data model storage unit 150, an attribute group accessed simultaneously based on an access pattern to the attribute of the entity is extracted, hierarchized, and the physical data model storage unit XML document format design means 123 for updating contents;
An XML schema generation unit 124 that acquires a physical data model from the physical data model storage unit 150, creates an XML schema corresponding to the physical data model, and stores the XML schema in the XML schema storage unit 160.

The XML collection design means 122 of the present invention (Claim 2)
From the access pattern of the access pattern storage means 140, either the basic format or the integrated format based on either the access frequency to each entity for each application, the frequency of reference or update, or the number of target data that meets the conditions Means for selecting.

The XML document format design means 123 of the present invention (Claim 3)
The extracted attribute group is grouped according to the access pattern to the attribute.

The XML document format design means 123 of the present invention (Claim 4)
The attribute group that is accessed simultaneously from the attribute access pattern is extracted, and the attribute group extracted from one application conflicts with or overlaps with other applications, and is hierarchical when there is a complete inclusion relationship Includes means for grouping and setting the group name.

  FIG. 2 is a diagram for explaining the principle of the present invention.

The present invention (Claim 5) is a data modeling method for designing an XML data structure of a database (XMLDB) connected to a user terminal and storing XML (Extensible Markup Language) of a structured document,
XML definition information storage means for storing logical XML definition information, access pattern storage means, physical data model storage means, XML schema storage means , logical design model design means, access pattern input means, and XML collection design means And an apparatus having XML document format design means and XML schema generation means ,
The logical model design means defines logical XML corresponding to each entity of the actual relationship diagram (ER diagram) representing the logical data model, and stores the logical XML definition information in the logical XML definition information storage means together with the ER diagram information. Model design step (Step 1),
An access pattern input step (step 2) in which the access pattern input means acquires the access pattern of the input application and stores it in the access pattern storage means;
When a set of entities having a relationship is specified from the ER diagram , the XML collection design means is individually set for each entity based on the logical XML definition information of the logical XML definition information storage means and the access pattern of the access pattern storage means. When update processing is accessed, the entity is represented as a basic format that expresses a relationship using a foreign key. When multiple entities are accessed together, the entity is physically represented as an integrated format that represents a tree structure that integrates multiple entities. XML collection design step (step 3) for storing in the data model storage means;
The XML document format design unit obtains an entity from the physical data model storage unit, extracts attribute groups that are accessed simultaneously based on the access pattern to the attribute of the entity, performs stratification, and forms the physical data model. XML document format design step (step 4) for updating the contents of the physical data model storage means;
An XML schema generation means, and acquires the physical data model from a physical data model storage unit, to create an XML schema corresponding to the physical data model, XML schema generation step of storing the XML schema storing means (step 5), I do.

Further, the present invention (Claim 6), in the XML collection design step,
Based on either the access frequency of each entity for each application, the frequency of reference or update, or the number of target data that meets the conditions, from the access pattern of the access pattern storage means, either basic format or integrated format select.

In the present invention (Claim 7), the attribute group extracted in the XML document format design step is grouped according to an access pattern to the attribute.

In the XML document format design step, the present invention (claim 8)
The attribute group that is accessed simultaneously from the attribute access pattern is extracted, and the attribute group extracted from one application conflicts with or overlaps with other applications, and is hierarchical when there is a complete inclusion relationship Perform grouping and set the group name.

  The present invention (Claim 9) is a data modeling program for causing a computer to function as each means constituting the data modeling apparatus according to any one of Claims 1 to 4.

  As described above, according to the present invention, in data modeling for XMLDB, conceptual design applies the conventional method for relational database (RDB) as it is, and logical design uses the expressive power of XML tree structure model. In the physical design, the relationship between entities is expressed by the primary key and foreign key attributes, and the expression by the structure in one XML document. By selecting the format (basic format / integrated format), the performance can be improved while maintaining the normal form. Furthermore, a highly convenient data structure can be designed.

It is a principle block diagram of this invention. It is a figure for demonstrating the principle of this invention. It is a figure which shows the comparison of the method of this invention, and the conventional method. It is a block diagram of the modeling apparatus in one embodiment of this invention. It is an example of expression of M: N relation by a tree structure model in one embodiment of the present invention. It is an example of expression of two kinds of relationships in one embodiment of the present invention. It is a figure which shows the expansion (notation of an integrated format) of the notation of the ER figure in one embodiment of this invention. It is a flowchart of operation | movement of the XML collection design part in one embodiment of this invention. It is a figure which shows the correspondence of the notation of grouping and XML expression in one embodiment of this invention. It is a screen image of the modeling apparatus in one Example of this invention. It is an example of logic XML of one Example of this invention. It is an example of the access pattern to the entity of one Example of this invention. It is an example of the access pattern to the attribute of one Example of this invention. It is an example of ER figure after the XML collection design of one Example of this invention. It is an example of the ER figure of the physical data model after grouping of one Example of this invention. It is an example (XML collection and its data structure) of the XML data of one Example of this invention. It is an example of the XML schema of one Example of this invention. It is an example of the ER figure of a conceptual data model. It is an example of a UML class diagram. It is an example of the ER figure of the logical data model for RDB. It is an example of the ER figure of the physical data model for RDB. It is an example of XML data and its structure.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  Since the conceptual design is independent of DBMS, a conventional method is used as it is, and a method of logical design and physical design is proposed. In particular, in physical design, we propose a method for maintaining both normal form and improving performance.

  In addition, an automated method using support tools is also proposed.

  First, FIG. 3 shows an overall procedure compared with that for a conventional RDB. In the same figure, “design of logical XML”, “design of XML collection”, and “design of XML document format” surrounded by bold lines are the parts proposed by the present invention, and will be described below. Other parts are assumed to use the same method as in the prior art, and the description thereof is omitted.

  In the present invention, data modeling for XMLDB will be described.

  FIG. 4 shows the configuration of the modeling apparatus in one embodiment of the present invention.

  A modeling apparatus 100 shown in the figure is connected to a user terminal 300, and includes a concept / logical model management unit 110, a physical model management unit 120, a concept / logical model data (logical XML definition information) storage unit 130, and an access pattern storage. Unit 140, physical model data storage unit 150, and XML schema storage unit 160.

  The physical model management unit 120 includes an access pattern input unit 121, an XML collection design unit 122, an XML document format design unit 123, and an XML schema generation unit 124.

  The concept / logical model management unit 110 stores the ER diagram and UML class diagram drawn by the user on the terminal 300 in the concept / logical model data storage unit 130 as concept / logical model data (logical XML definition information).

  The access pattern input unit 121 of the physical model management unit 120 stores the access pattern input by the user at the terminal 300 in the access pattern storage unit 140. However, the input may be in a format that passes a CSV (Comma Separated Values) file.

  The XML collection design unit 122 of the physical model management unit 120 refers to the information in the concept / logical model data storage unit 130 and the access pattern storage unit 140 when the user selects execution (presses a button, etc.) from the terminal 300, and the XML collection (XML document type) Design. At this time, interaction with the user is included as necessary. The XML collection design result is stored in the physical model data storage unit 150 as physical model data.

  The XML document format design unit 123 of the physical model management unit 120 designs the XML document format when the user directly designates (draws) grouping of attribute groups on the terminal 300 and selects execution (operation such as pressing a button). Then, based on the data in the access pattern storage unit 140 and the physical model data storage unit 150, the parts not directly designated by the user are grouped. When performing grouping, interaction with the terminal 300 is included as necessary. The design result is stored (updated) in the physical model data storage unit 150.

  When the user selects execution (operation such as pressing a button) from the terminal 300, the XML schema generation unit 124 of the physical model management unit 120 creates an XML schema based on the data in the physical model data storage unit 150, Stored in the schema storage unit 160.

  In the following, the logical design for XMLDB and the physical design for ML will be described in the order of FIG.

<Logical design> (Logical XML design)
The logical design is performed by the concept / logical model management unit 110.

  First, in the logical design for RDB, transformation of the ER diagram (normalization to the relational model) was required due to restrictions on the relational model, but in the case of XMLDB, the expressive power of the tree structure model is high. Therefore, no deformation is required. Specifically, the M: N relationship can also express the foreign key attribute as a repeated item.

  For example, the relationship between the club activities shown in FIG. 18 and the student's M: N can be expressed as a data structure of a tree structure model as shown in FIG. An intermediate node “club activity list” is created in the student entity, and a plurality of external keys “club activity codes (FK)” corresponding to the primary key of club activities are held as repeated elements.

  Next, the concept / logical model management unit 110 defines “logical XML” corresponding to the ER diagram and stores it in the concept / logical model data storage unit 130. In the case of RDB, an entity corresponds to a logical table, an attribute corresponds to a logical column, and a relationship corresponds to a logical column as a foreign key attribute. In the case of logical XML, the correspondence is as follows.

  Entity: Corresponds to one logical XML.

・ Attribute: Corresponds to a node in logical XML (whether it is expressed as an element or XML attribute is not specified)
・ Relationship: The expression format is not stipulated (the specific data design of the relationship is done by physical design, so no work beyond holding information is done here).

  Therefore, the support tool of the conventional data modeling technique can also be used as it is (it is sufficient if the conceptual data model can be drawn).

<Physical design>
Next, physical design for XMLDB in the physical model management unit 120 will be described. First, the basic concept will be described.

  ・ Relationship can be selected from two types (basic format and integrated format).

-Add primary key and foreign key attributes (same method as RDB);
-Representation by tree structure in XML document (one entity does not have to be one kind of XML document. Multiple entities can be represented by one kind of XML document).

  -Attributes in entities can be layered by adding intermediate nodes.

  An example of the two types of expression methods of the former relationship is shown in FIG. The left side is a conventional foreign key representation (basic form), which is divided into two types of XML documents for each entity, and the relationship is maintained by “student ID”. The right side is an example of the XML tree structure representation (integrated format). The “student” entity and the “exam” entity are integrated into one type of XML document, and expressed in the relationship between the parent and child nodes of the tree structure. Has been. The “test result” entity is a repeated item because “student” and “test result” have a 1: N relationship.

  As described above, there are two types of relationship representation methods, and the expression by the foreign key is hereinafter referred to as “basic form”, and the expression by the tree structure is hereinafter referred to as “integrated form”. In addition, the type of XML document is hereinafter referred to as “XML collection”. That is, two entities having a relationship become two XML collections in the basic format (left side in FIG. 6), and correspond to one XML collection in the integrated format (right side in FIG. 6).

  In the integration format, basically, a parent entity on the “1” side of 1: N can be centered and a child entity on the “N” side can be expressed as a repetition node. Further, if there is a 1: N relationship with another entity for a child entity, the integration format can be applied in a nested manner. Also in this case, it is important that the data is not redundant and can be expressed without breaking the normal form.

  However, it is not unlimited and is in a range that satisfies the following conditions, but it is in a wider range compared to the conventional RDB.

[conditions]
-When a child entity has multiple parent entities, only one parent is expressed in an integrated format, and the remaining relationships are expressed in a basic format.

  -M: N relationship is expressed in a basic format (not expressed in an integrated format).

  For the following explanation, the notation of the ER diagram is expanded, and as shown in FIG. 7, the entity group (strictly, the relationship group) surrounded by a dotted line represents the application of the integrated format.

  In the support tool for modeling technique, it is necessary to prepare a new notation in an integrated format. Of course, in addition to enclosing with dotted lines, various other notations such as changing the format of the relationship lines can be considered.

<XML collection design>
Next, processing of the XML collection design unit 122 will be described.

  It was mentioned above that an XML collection is determined by a relationship representation. An XML collection corresponds to an RDB table, and is designed from the viewpoint of improving performance, as in the case of RDB.

  The performance characteristics that can be expected differ between the basic format and the integrated format, and are generally as follows.

・ When the basic type is advantageous in performance:
When individual update access is made for each entity. In particular, it is advantageous when inserting or deleting a parent entity, inserting a child entity, or updating a foreign key.

・ If the integrated format offers performance advantages:
When both entities are accessed together. It is advantageous at the time of combined search (reference), batch insertion, and batch deletion.

  In other words, since the integrated form is managed in a state where two entities are grouped together, it is advantageous to have a grouped access. On the other hand, when accessing the update system, one change does not have to be done alone, the other change The basic form is advantageous.

  Accordingly, the XML collection design unit 122 selects and designs an XML collection in consideration of the behavior such as how and how often the business application (AP) accesses which entity.

  An arrangement of the AP behavior is referred to herein as an “entity access pattern”, which is created by the access pattern input unit 121 in the “arrangement of access patterns” work before the XML collection design work, and an access pattern storage unit 140.

Organization of access patterns is done in the same way in the conventional method,
-Washing out AP;
-Access frequency to each entity for each AP, reference / update, number of hits (number of target data that matches the conditions)
・ Information such as the number and distribution of data for each entity is organized.

  FIG. 8 shows an XML collection design procedure performed by inputting an ER diagram of a logical data model and an access pattern to an entity. This can also be (semi) automated by support tools.

  The operation of FIG. 8 will be described below.

  When the system scans all the relationships at the timing when the “XML collection design” button of the screen image is pressed by the user, the XML collection design unit 122 creates a set of entities having relationships from the physical data model storage unit 150. Select (step 101). The XML collection design unit 122 refers to the concept / physical model data storage unit 130 to determine whether the set of entities selected in step 101 is in the M: N relationship, and if M: N (step 102, Yes). ) "Basic format" is selected (step 107). If not in the M: N relationship, the access pattern storage unit 140 is referred to determine whether there is an AP to be accessed collectively (step 102). If not (No in step 103), “basic format” is selected. (Step 107). On the other hand, if there are APs that are collectively accessed (step 103, Yes), it is determined whether there are APs that are individually accessed for the update system (step 104, No), and “integrated format” is selected. (Step 106). If there is (Step 104, Yes), both AP groups are comprehensively judged. If it can be estimated that the integration is advantageous (Step 105, Yes), the “integration format” is selected (Step 106). When it is estimated that it is not advantageous (No at Step 105), "basic form" is selected (Step 107).

  In the above-mentioned step 105, as the “method for comprehensively determining the AP group”, as a method for automatically determining, the ratio between the access frequency and the update / reference cost is determined in advance, and the access frequency × There may be a method of selecting a format in which the higher total cost is advantageous.

  If the above processing is performed for all entity pairs (step 108, Yes), the most advantageous one is the relationship when there are relationships to a plurality of parent entities in the range where "integrated form" is selected. The parent entity is selected, and the others are changed to “basic format” (step 109).

  When the AP group that accesses the two entities collectively and the AP group that individually accesses the update system compete with each other by referring to the access pattern storage unit 140 with reference to the access pattern storage unit 140, the performance of both groups is improved. It is necessary to examine the influence of the At this time, the access frequency is an important index, but in addition, general access pattern information such as the number of data and the number of hits (the number of target data that matches the conditions) should be considered. Note that the access pattern storage unit 140 is referred to for the access frequency. Depending on the frequency, the access frequency may be expressed as a level such as high, medium, or low, or may be expressed as the number of accesses.

  Further, in step 109 described above, since integration into a plurality of parent entities breaks the normal form, it is necessary to select one parent entity. As in step 105, this also requires an integrated study from the access frequency.

As a (semi) automated method in the support tool in steps 105 and 109,
・ A method to determine specific formulas and rules with various indicators as input and to make judgments automatically;
-A method of presenting a group of indicators, calculation results, etc. to the user (designer) and requesting judgment from the user;
Can be considered.

  By using the above procedure, an XML collection is designed in the “integrated format” for those that are accessed by multiple entities from the AP in a unified manner, and an XML collection is designed in the “basic format” for those in which individual update access is performed. In addition, since the normal form of the data model can be maintained, both performance improvement and maintainability can be achieved.

  In addition, if it is outside the scope of the present invention and priority is given to further improvement in performance, it is possible to design with the normal form destroyed in the same manner as in the case of the conventional RDB in addition to the above.

<Design of XML document format>
Next, an XML document format design process performed by the XML document format design unit 123 will be described.

  In order to make the data structure more convenient by utilizing the expressive power of the XML tree structure model, the XML document format design unit 123 stratifies the attributes in the entity. This is also determined from the access pattern of the AP in the access pattern storage unit 140, but is determined from the access pattern to the attribute in the entity, which is more detailed information. This is an arrangement of behaviors such as how often each AP accesses which attribute for each entity. This is also done conventionally by the “organization of access patterns” operation, and the APs accessed for each attribute, the type of reference / update, etc. are organized and stored in the access pattern storage unit 140.

  The design procedure for the XML document format is shown below. The XML document format design unit 123 performs the following processing for each entity.

  (1) The XML document format design unit 123 refers to the access pattern storage unit 140, and extracts attribute groups that are accessed simultaneously from the access pattern to the attributes.

  (2) If the attribute group extracted from the AP of (1) above does not conflict with those of other APs, the XML document format design unit 123 performs grouping as it is (group name setting and belonging attribute group) Decision).

If the extracted attribute group ranges conflict / overlap,
・ If there is a complete inclusion relationship, perform hierarchical grouping;
・ Prefer insert / update / delete over reference;
・ Consider priority on processing costs such as access frequency;
From such a viewpoint, grouping is performed comprehensively and a group name is set.

  (3) In the internal processing at the AP, if a highly convenient structure is clear, further grouping is performed. Here, when it is determined whether or not a highly convenient structure is clear, for example, when “address”, “name”, “age”, and “profession” are displayed on the screen by the AP, “name” and “age” are changed. When the processing is different, such as displaying like a headline and displaying the others small, it can be determined that the convenience is high if they are grouped.

  At this time, the group name should be different from all entity names and attribute names in the entity.

  FIG. 9 shows an expanded ER diagram notation (left side) and a data structure example of XML representation (right side) when grouped. In the ER diagram notation, the XML document format design unit 123 extends the notation so that the group name description of grouping and the attributes belonging to the group are expressed by indentation.

  As a (semi) automated method using the support tool, a notation method representing grouping as shown in FIG. 9 is added. However, since it is difficult to fully automate the above design procedure, various indicators are presented on the terminal 300 of the designer. Then, a method for requesting judgment from the user can be considered.

  In the XML representation, it is natural to represent it as an intermediate node (an intermediate element that is neither the root element nor the lowest layer element) with the group name as the element name.

  In conventional RDB, by issuing a query (SQL “SELECT” clause) that lists and specifies only the items required by the AP (attributes in the entity), the amount of data transfer is minimized and performance is improved. It has been done. In XMLDB, the same method can be used to improve performance, but by taking advantage of the tree structure data model and performing the above grouping, group names can be specified instead of listing attributes here. , Not only improves performance, but also improves maintainability by simplifying queries.

<Create XML schema>
Finally, an XML schema creation process in the XML schema creation unit 124 will be described.

Specifies the final XML representation corresponding to the ER diagram. Insufficient provision of XML collection for XML documents and logical XML rules. ・ Representation of relationships. Specifically, the basic form of key attribute representation and the integrated form of structural representation:
M: N-related foreign key expression:
Determine.

  Based on this, an XML schema corresponding to the physical data model in the physical model data storage unit 150 is created and stored in the XML schema storage unit 160.

  Similar to the creation of the conventional RDB schema, the XML schema can be automatically generated by the support tool from the definition of the XML expression.

  The present invention itself does not particularly define a definition for a specific XML representation. All information should be stored, and expressions that are easy to access and operate should be considered and determined.

  Hereinafter, as an example, a design example based on the ER diagram for the conceptual data model of FIG. 18 is shown.

  FIG. 10 is a screen image of the modeling apparatus according to the embodiment of the present invention.

<Logical design> (Logical XML design)
First, the logic design in the concept / logic model management unit 110 will be described.

  In the logic design, the ER diagram is not particularly modified.

  The definition of logical XML for the representation of ER diagram is defined as follows.

·entity:
The root element is an entity element, and the element name (tag name) is an entity name.

·attribute:
The element immediately below the entity element. The element name (tag name) is the attribute name, and the value is stored in the text node.

  FIG. 11 is an example of the logical XML of one embodiment of the present invention, and shows the logical XML corresponding to the ER diagram of FIG. 18 based on the above definition. In the figure, the relationship is omitted.

  In the screen image shown in FIG. 10, when the “concept / logical model” tab is selected by the user, an ER diagram can be drawn, and the definition information of the logical XML is held in the conceptual / logical model data storage unit 130. It shall be.

<Physical design>
First, the access pattern input unit 121 of the physical model management unit 120 acquires an access pattern to the entity shown in FIG.

  In this example, five APs are plotted on the vertical axis, and access to each entity is summarized in a table format. The access frequency (high / medium / low) for each AP and the type of access are C (new creation), B (reference), U (update), and D (delete).

  Further, it is assumed that the pattern shown in FIG. 13 is obtained as the attribute access pattern of the student entity. In the example shown in the figure, the four APs that access the student entity have the same table format on the vertical axis.

  In the screen image shown in FIG. 10, the “physical model” tab is selected, the “input access pattern” button is pressed, and the access pattern is input. As a result, the access pattern input unit 121 stores the access pattern to the entity shown in FIG. 12 and the access pattern of the attribute shown in FIG. 13 in the access pattern storage unit 140.

<XML collection design>
In the following, an example is shown in which the XML collection design unit 122 performs design according to the procedure shown in FIG. 8 from the access pattern shown in FIG. 12 stored in the access pattern storage unit 140 with respect to the ER diagram of FIG. .

  Further, it is assumed that the determination at the time of contention is made with reference to the access frequency of the AP stored in the access pattern storage unit 140.

[1] "Basic format" selection:
When the group of “class (parent) and student (child)” is selected in step 101 of FIG. 8, it is not the relationship of M: N (step 102, No), but it is referred collectively by the class list acquisition AP. (Step 103, Yes), only the student (child) is updated by the student information change AP (Step 104, Yes). Since the access frequency is such that the class list acquisition AP <student information change AP (step 105, No), “basic format” is selected (step 107).

[2] "Basic format" selection:
When the group of “club activity (parent) and student (child)” is selected in step 101 of FIG. 8, the “basic form” is selected because of the M: N relationship (step 102, Yes) (step 107). ).

[3] "Integrated format" selection:
If the pair of “student (parent) and exam grade (child)” is selected in step 101 of FIG. 9, the student's first name / first name and exam results are acquired instead of the M: N relationship (No in step 102). AP is collectively referred to (step 103, Yes), only the student (parent) is updated by the student information change AP, and only the test result (child) is created by the registered AP of the test result. (Step 104, Yes). Since the access frequency is “student's first and last name and test result acquisition AP> student information change AP + test result registration AP” (step 105, Yes), “integrated form” is selected (step 106).

  In step 108, it is assumed that all the processes have been performed, and there is no selection of “integrated form” for a plurality of parents.

  By pressing the “XML collection design” button on the screen image of the modeling apparatus shown in FIG. 10, the above result can be automatically reflected in the ER diagram drawing. Therefore, the ER diagram is as shown in FIG. “Student” and “Examination results” are surrounded by a dotted line of “integrated form”, and “basic form” foreign keys of “class” and “club activity” are added to “student”.

<Design of XML document format>
Next, a specific example of the XML document format design unit 123 is shown below.

  From the access pattern of the access pattern storage unit 140 shown in FIG. 13, it can be seen that the “student” entity refers to the last name and the first name simultaneously from three APs. Therefore, the XML document format design unit 123 groups the first name and last name by “name” and stores them in the physical model data storage unit 150.

  For example, when the user selects the “Student” entity in the screen image shown in FIG. 10 (FIG. 10) and presses the “XML document format design” button, the XML document format design unit 123 stores the physical model data storage unit 150. The access pattern for each attribute of the “student” entity can be displayed on the terminal 300 to assist in the determination made by the user.

  FIG. 15 shows an ER diagram of the final data model after the XML document format design unit 123 performs grouping for the example of FIG.

<Generate XML schema>
Finally, a specific example of the XML schema generation unit 124 is shown below.

The correspondence of XML expressions to ER diagrams is specified as follows. In addition to the provisions of logical XML,
-The entity name that is the first parent in the XML collection is the root element name;
・ For the foreign key attribute in the basic format, the element name (tag name) is used as the attribute name and the value is stored in the text node as in the normal attribute;
-An integrated child entity creates an intermediate node named "entity name +" list "" directly below the parent entity element, and makes it a repeated element;
For the M: N related foreign key attribute, create an intermediate node named “attribute name +“ list ”” and express the foreign key attribute as a repeating element;
Normally, a physical name different from the logical name is defined for various names, and a physical name is assigned to final data. In this embodiment, the physical name and the logical name are set to be the same.

  FIG. 16 shows an image of XML data (data structure of three XML collections) to which the above rules are applied.

  It is assumed that the above rules are stored in the physical model data storage unit 150. In the support tool (FIG. 10), when the “create schema” button is pressed from the terminal 300, the logical XML of the physical model data storage unit 150 is also displayed. An XML schema (description example based on XML Schema) as shown in FIG. 17 can be automatically generated from the included rules.

  Note that the operations of the concept / logical model management unit 110 and the physical model management unit 120 described above are constructed as programs and can be installed in a computer used as a data modeling apparatus and executed, or distributed via a network. Is possible.

  Further, the constructed program can be stored in a portable storage medium such as a hard disk, a flexible disk, or a CD-ROM, and can be installed or distributed in a computer.

  The present invention is not limited to the above-described embodiments and examples, and various modifications and applications can be made within the scope of the claims.

100 Modeling Device 110 Concept / Logical Model Design Unit, Concept / Logical Model Management Unit 120 Physical Model Management Unit 121 Access Pattern Input Unit, Access Pattern Input Unit 122 XML Collection Design Unit, XML Collection Design Unit 123 XML Document Format Design Unit, XML Document format design unit 124 XML schema generation unit, XML schema generation unit 130 Logical XML definition information storage unit, concept / logical model data storage unit 140 Access pattern storage unit, access pattern storage unit 150 Physical data model storage unit, physical data model storage Unit 160 XML schema storage means, XML schema storage unit 300 terminal

Claims (9)

A data modeling device for designing an XML data structure of a database (XMLDB) connected to a user terminal and storing XML (Extensible Markup Language) of a structured document,
A logical model design unit that defines a logical XML corresponding to each entity of an actual relationship diagram (ER diagram) representing a logical data model, and stores the logical XML definition information in the logical XML definition information storage unit together with the information of the ER diagram as logical XML definition information;
An access pattern input means for acquiring an access pattern of the input application and storing it in the access pattern storage means;
When a set of entities having a relationship is specified from the ER diagram, an individual update process is performed for each entity based on the logical XML definition information of the logical XML definition information storage unit and the access pattern of the access pattern storage unit The physical data model represents the entity as a basic format that expresses the relationship with a foreign key when multiple accesses are made, and as an integrated format that represents a tree structure that integrates multiple entities when multiple entities are accessed together XML collection design means for storing in the storage means;
An entity is acquired from the physical data model storage means, an attribute group that is accessed simultaneously based on an access pattern to the attribute of the entity is extracted, hierarchized, and the physical data model storage means XML document format design means for updating content,
XML schema generation means for acquiring the physical data model from the physical data model storage means and creating an XML schema corresponding to the physical data model;
A data modeling apparatus comprising: The XML collection design means is:
Based on the access pattern of the access pattern storage means, the frequency of access to each entity for each application, the frequency of reference or update, or the number of target data that matches the conditions, the basic format or the integrated format The data modeling apparatus according to claim 1, further comprising means for selecting one of them. The XML document format design means includes:
The data modeling apparatus according to claim 1, further comprising means for grouping the extracted attribute group according to an access pattern to the attribute. The XML document format design means includes:
Extract attribute groups that are accessed simultaneously from the access pattern to the attribute, and if the attribute group extracted from one application conflicts or overlaps with other applications and has a complete inclusion relationship, it is hierarchical 4. The data modeling apparatus according to claim 3, further comprising means for performing grouping and setting a group name. A data modeling method for designing an XML data structure of a database (XMLDB) connected to a user terminal and storing XML (Extensible Markup Language) of a structured document,
XML definition information storage means for storing logical XML definition information, access pattern storage means, physical data model storage means, XML schema storage means , logical design model design means, access pattern input means, and XML collection design means And an apparatus having XML document format design means and XML schema generation means ,
The logical model design means defines a logical XML corresponding to each entity of an actual relationship diagram (ER diagram) representing a logical data model, and stores it in the logical XML definition information storage unit as information on the ER diagram as logical XML definition information. Logical model design steps to
The access pattern input means, an access pattern input storing acquires the access patterns of the input application to the access pattern storage means,
When the XML collection design means designates a set of entities having a relationship from the ER diagram, based on the logical XML definition information of the logical XML definition information storage means and the access pattern of the access pattern storage means, When access to individual update processing is made for each entity, the basic format is used to express the relationship using a foreign key, and when multiple entities are accessed together, integration is expressed using a tree structure that integrates multiple entities. XML collection design step for storing entities in physical data model storage means as a format;
The XML document format design means acquires an entity from the physical data model storage means, extracts an attribute group that is accessed simultaneously based on an access pattern to the attribute of the entity, performs stratification, and a physical data model XML document format design step for updating the contents of the physical data model storage means as
The XML schema generating means acquires the physical data model from the physical data model storage means, creates an XML schema corresponding to the physical data model, and stores the XML schema in the XML schema storage means;
A data modeling method characterized by: In the XML collection design step,
Based on the access pattern of the access pattern storage means, the frequency of access to each entity for each application, the frequency of reference or update, or the number of target data that matches the conditions, the basic format or the integrated format 6. The data modeling method according to claim 5, wherein either one is selected. In the XML document format design step,
The data modeling method according to claim 5, wherein the extracted attribute group is grouped according to an access pattern to the attribute. In the XML document format design step,
Extract attribute groups that are accessed simultaneously from the access pattern to the attribute, and if the attribute group extracted from one application conflicts or overlaps with other applications and has a complete inclusion relationship, it is hierarchical 8. The data modeling method according to claim 7, wherein grouping is performed and a group name is set.   A data modeling program for causing a computer to function as each means constituting the data modeling apparatus according to any one of claims 1 to 4.

Images