KR100456027B1 - Apparatus and method for transforming ebXML filter query by using G-FQL - Google Patents

Abstract

The present invention relates to an apparatus and method for converting an ebXML filter query using G-FQL. After the G-FQL converter 10 converts an ebXML filter query to a graph model based query word defined in XML syntax, G-FQL, After the XML-object converter 20 converts the G-FQL of the XML syntax into a user-operable program object, the query graph generator 30 has a meaning equivalent to that of the G-FQL from the G-FQL converted to the object. Then, after the SQL generator 40 optimizes the query word in the query graph to generate the SQL, the query executor 50 performs the query using the optimized SQL and retrieves the object. The XML converter 60 is configured to convert the retrieved object back to XML,

Accordingly, when the ebXML filter query is modified or extended, it has the advantage of minimizing the update of the query processing device of the database management system. The ebXML filter query can be generalized by converting the ebXML filter query into various database queries such as OQL or XQuery as well as SQL. There is an advantage.

Description

Apparatus and method for transforming ebXML filter query by using G-FQL}

The present invention relates to a query processing apparatus of a database management system (DBMS), and more particularly, to a filter query provided in an e-business eXtensible Markup Language (ebXML) Registry using a Generalized Filter Query Language (G-FQL). The present invention relates to an apparatus and method for converting an ebXML filter query using G-FQL that converts a language into various database query languages such as SQL (Structured Query Language), OQL (Object Query Language) or XQuery.

In a query processing apparatus of a conventional database management system, before processing a query, it is necessary to first determine where to store an ebXML Registry Information Model (ebRIM) object, and the existing object and relational relation as a storage medium of the ebRIM. ) Or an XML database system.

In addition, the ebXML filter query is defined in XML syntax for ebRIM object search stored in the ebXML registry as shown in FIG. 1, and one filter query corresponds to one class defined in ebRIM. Is a set of class instances.

The ebRIM defines the metadata required to describe business information. It refers to the classes defined in ebRIM, and uses a filter query to define a constraint on one metadata and the relationship between multiple metadata. ) Can be used to query.

In the ebXML filter query, the restriction is represented by a filter element and the relationship is represented by the name of a query or branch element defined in the filter query.

Here, the important point is that the element name implies the relationship between ebRIM objects in the ebXML filter query, and this characteristic can be seen in the example of 'RegisryObjectQuery' in the ebXML filter query of FIG. If you look at the <classifiedByBranch> element of 1, there is no attribute or subElement that represents the relationship between ebRIM objects, but it implies the relationship between the 'RegistryObject' and 'Classification' objects of ebRIM.

However, conventional database systems do not support the ebXML filter query and can query ebRIM objects stored in the database only through database queries such as SQL, OQL, or XQuery.

Therefore, in order to retrieve ebRIM objects stored in relational database using ebXML filter query, the conversion process of converting ebXML filter query to SQL is necessary. If ebRIM is stored in object database, it must be converted to OQL.

In particular, if the user does not understand the meaning of the relationship between the ebRIM objects implied by the element names as described above, the query cannot be used. Furthermore, the 11 query elements defined in the well-known ebXML filter query and these It is difficult to apply the general processing method of each query element and branch element for query transformation because the relationship between the ebRIM objects contained in each of the branch elements that can be positioned as subelements of is different.

In fact, in the existing database management system, since the ebXML filter query is dependent on a specific information model called ebRIM, the process of converting to a query for a database or XML document such as SQL, OQL or XQuery can be generalized. There is no problem.

Accordingly, an object of the present invention is to overcome the above-described problems, and an object of the present invention is to convert a ebXML filter query into a query language provided by a database management system, or a relational model, an object model, or a semi-structured object. Using G-FQL which is based on graph model that can express query about various data models such as model and converts filter query provided by ebXML registry to SQL using G-FQL which is a query defined by XML syntax. An apparatus and method for converting ebXML filter quality are provided.

An apparatus for converting an ebXML filter query using G-FQL for achieving the object of the present invention includes: a G-FQL converter for converting a user-defined ebXML filter query into a G-FQL, which is a query model based on a graph model defined by XML syntax; ; An XML-object converter that converts G-FQL of XML syntax into a user-operable program object; A query graph generator for generating a query graph having a meaning equivalent to G-FQL from the G-FQL converted to the object; An SQL generator for generating SQL by optimizing a query in the query graph; A query processor that executes a query by using the optimized SQL generated; And an object-XML converter for converting the object retrieved by the query executor back into XML.

The method for converting an ebXML filter query using G-FQL for achieving the object of the present invention includes the steps of: converting a user-defined ebXML filter query to a G-FQL, which is a graph model based query word defined in XML syntax; Converting the G-FQL of the XML syntax into a program object operable by the user and generating a query graph having the same meaning as the G-FQL from the G-FQL converted into the object; Generating SQL by optimizing a query in the query graph; And retrieving the object by converting the retrieved object back into XML by performing a query using the optimized generated SQL.

1 is an embodiment illustrating an ebXML filter query;

2 illustrates an apparatus for converting ebXML filter quality using G-FQL according to the present invention.

Diagram,

3 illustrates a method of converting ebXML filter quality using G-FQL according to the present invention.

Flowchart,

4 illustrates an example of XSL for converting an ebXML filter query to G-FQL.

5A illustrates an embodiment of an XML schema syntax defining a structure of a G-FQL.

5B shows that the ebXML filter query shown in FIG. 1 is converted by the XSL shown in FIG.

Example showing the G-FQL,

Figure 6a is a general embodiment of the G-FQL model,

FIG. 6B is a graph showing the G-FQL shown in FIG. 5B;

FIG. 7A illustrates an embodiment in which the G-FQL query graph shown in FIG. 6B is converted into SQL. FIG.

FIG. 7B illustrates an embodiment in which the G-FQL query graph shown in FIG. 6B is converted into optimized SQL.

to be.

<Description of the symbols for the main parts of the drawings>

10: G-FGL Converter 20: XML-Object Converter

30: Query Graph Generator 40: SQL Generator

50: Query Performer 60: Object-XML Converter

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Referring to FIG. 2, the G-FQL converter 10 converts a user-defined ebXML filter query to G-FQL, which is a graph model based query defined by XML syntax.

The G-FQL converter 10 uses G-FQL, which is a query term based on a graph model and defined in XML syntax, as an intermediate query word for converting an ebXML filter query into a database query such as SQL, OQL, or XQuery. This is because graph models can express queries on various data models supported by the database management system, such as relational models, object models, or semi-structured models, and can be easily converted to SQL or OQL. In fact, when the ebXML filter query is converted to G-FQL by the G-FQL converter 10, the ebXML filter query is directly converted to a database query such as SQL without going through the process of converting the ebXML filter query to G-FQL. In this case, among the program modules for query conversion, there are many factors that depend on the type of data model in which ebRIM is stored. As a result, when the ebXML filter query is added or changed, the change of the query conversion device becomes complicated.

The G-FQL expresses the relations implied by the names of the query or branch elements defined in the ebXML filter query, and expresses the relations explicitly using a graph model to express various relationships defined in the ebXML filter query. The meanings of two queries or branch elements can be expressed in the same model, and by using this model, the ebXML filter query can be generalized and converted into various database queries such as SQL, OQL, and XQuery.

The XML-object converter 20 converts the G-FQL of the XML syntax into a user-operable program object.

The query graph generator 30 generates a query graph having the same meaning as the G-FQL from the G-FQL converted into an object.

The SQL generator 40 generates SQL by optimizing a query in the query graph.

The query executor 50 searches for an object for the query by performing a query using the optimized generated SQL.

The object-XML converter 60 converts the object retrieved by the query executor back into XML.

An apparatus for converting ebXML filter quality using G-FQL according to the present invention configured as described above operates as follows.

Referring to FIG. 3, first, as the user defines an ebXML filter query, the G-FQL converter 10 converts the user-defined ebXML filter query to G-FQL, which is a graph model based query defined by XML syntax. (S10).

In this case, the G-FQL converter 10 converts a user-defined ebXML filter query to G-FQL expressed in XML using a known eXtensible Style Language Transformations (XSLT) engine and XSL as shown in FIG. 4. In particular, in order to be able to convert all the filter queries provided by the ebXML filter query as shown in FIG. 1 to G-FQL, as shown in FIG. 4, all query elements and branch elements defined in the ebXML filter query are shown. Just add an application template.

In practice, the G-FQL converter 10 converts a user-defined ebXML filter query to G-FQL as shown in FIG. 1 using a known XSLT engine and XSL as shown in FIG. 4. The syntax is defined as an XML schema as shown in FIG. 5A and described as G-FQL, a query word having the same meaning as the filter query of FIG. 1 as shown in FIG. 5B. At this time, the G-FQL shown in FIG. 5B must follow the XML schema shown in FIG. 5A.

The G-FQL query expresses the relationship between the ebXML filter query, which contains the relationship between ebRIM objects in the element name, in order to convert the ebXML filter query into a general database query such as SQL, OQL, or XQuery as subelements and attributes. It is converted into the same meaning as the ebXML filter query.

The basic elements of the G-FQL query are a query element, a branch element, a subquery element, and a filter element, and each element will be described in detail as follows.

The query element is the top element that starts the query and only appears once in the query.

An attribute 'class' of the query element means an object to be queried and may have names of objects defined in ebRIM as a value. The attribute 'type' refers to a logical relationship between branch elements or elements of subqueries that can be placed under query elements. In the case of an ebXML filter query, the default value of the attribute 'type' is also logical because the logical operation between elements of the same level is implicitly always AND.

The query element may have a branch / subquery / filter as a sub element, and the query result depends on the attribute 'returnType'.

The possible value for attribute 'returnType' can be one of five attributes ('ObjectRef', 'RegistryObject', 'RegistryEntry', 'LeafClass', 'LeafClassWithRepositoryItem') defined in the ebXML Registry Services Specification. Reference to an object (ObjectRef).

The element of the subquery may have a filter / subquery / branch element as a sub element. The attributes 'class' and 'type' of the query element have the same meaning as defined in the query element.

The subquery contains a join operation with the target object of the parent element (the value of the attribute 'class' of the parent element). The value of the 'this' attribute of the subquery can be one of the attributes belonging to the target object of the current element, and the value of the attribute 'parent' is one of the attributes belonging to the target object of the parent element. . With the attributes 'this' and 'parent' the current element represents a join with the parent element.

Referring to the first branch element of FIG. 5B as an example, a join relationship is established between the property 'ClassifiedObject' of the 'Classification' object which is the target object of the branch element, and the property 'id' of the 'RegistryObject' object which is the target object of the parent element. That is, it indicates that Classificatioin.id = RegistryObject.id.

The branch element has the same meaning as the element of the subquery. In the case of an ebXML filter query, a subquery consists of several query elements and branch elements. In ebXML filter query, it is divided into branch or query element depending on whether it is reused or not. But in G-FQL, reusability doesn't mean much.

The reason for defining the same branch element as the subquery element in the G-FQL is simply to maintain the structure of the ebXML filter query.

The filter element is an element that defines a predisk for the target object to which the query / subquery / branch element is assigned to the attribute 'class'. The target object of the filter element is the same as the target object of the element immediately above it.

The target object of the first filter query of FIG. 5B is the same as 'Name' which is the target object of the upper branch element.

The filter element must contain one Clause element, and is identical to the Clause element defined in the ebXML Service Specification. That is, it refers to the Clause element defined in ebXML filter query. The ClassificatioinNodeFilter element and LocalizedStringFilter element processing portions of FIG. 4 also copy the following elements of the Clause element of FIG. 1 as they are.

The clause element defines a comparison operation and a logical operation on attribute values of the class instance targeted by the filter element. When a comparison and logical operation of a target object's property value is defined only once, it is called Simple Predicate, and if it is defined more than once, it is called Compound Predicate.

After converting the ebXML filter query to G-FQL, which is an intermediate query word for converting the general database query SQL into SQL, the XML-object converter 20 can manipulate the G-FQL expressed in XML syntax. After converting to an object, the query graph generator 30 generates a query graph having a meaning equivalent to that of G-FQL from the converted G-FQL to an object (S20).

In this case, the general form of the query graph generated from the G-FQL by the query graph generator 30 is represented as shown in FIG. 6A, and is actually equivalent to the G-FQL from the G-FQL shown in FIG. 5B. The query graph with is expressed more specifically as shown in FIG. 6B.

Referring to FIG. 6A, a general form of query graph according to the present invention is composed of two nodes and four arcs.

First, looking at the two vertices, the vertex [Root] means the starting point of the query graph. This is necessary to maintain a reference to the object to be queried. Vertex [Object] means the object to be queried and contains the identifier, type and attributes of the object.

The edge means a relationship between objects. Edges contain predicates that mean operations between the start and end object properties. According to the kind of predisk, it is divided into trunk <root>, trunk <filter>, trunk <join>, and trunk <query>.

The edge <root> is connected to one vertex [Object] among any vertex [Object] existing on the query graph starting from the vertex [Root]. Edge <root> contains the attribute 'returnType' and is the object whose vertex target object corresponding to the endpoint of edge <root> is returned as the result of the query.

The edge <join> performs a comparison operation (=, ~ =,>.> =, <, <=) On attributes included in any two vertices of the [Object] type present on the query graph. The two attributes of the comparison operator must be of the same data type.

The edge <filter> performs a comparison operation on the attributes of one vertex [Object].

The edge <query> means that there is a subquery relationship between two vertices [Object]. The result of the subquery performs the set operation with the target object of the starting vertex.

Referring to FIG. 6B, in the case of the ebXML filter query, an upper element target object and a lower element target object have attribute values of certain attributes of an upper element target object having the same data type and attribute values of some attributes of a lower element target object. Always include a membership operation (IS IN, IS NOT IN) for.

The ebXML filter query is implicitly defined to mean only IS IN operations between attributes of parent / sub elements. Therefore, the IS IN operation can be replaced by an equal comparison operation of two attribute values, and when converting the ebXML filter query to G-FQL, the edge <query> always includes only the equal comparison operation.

Therefore, the edge <join> is not used when the ebXML filter query expresses the relationship between the upper element and the lower element in the query graph. This is because ebXML filter queries always have a hierarchical structure of parent and child elements. This is because the definition of edges is necessary for the query graph to represent a general query.

After the query graph is generated as described above, the SQL generator 40 generates the SQL by optimizing the query word in the query graph (S30).

In this case, the SQL generated by optimizing the query in the query graph by the SQL generator 40 is represented as shown in Figure 7a.

The SQL shown in FIG. 7A includes the four subqueries shown in FIG. 6B as they are. The vertices of FIG. 6B correspond to table names located in the FROM clause of the SQL statement, the filter corresponds to the predicate of the WHERE clause, and the subquery is included in the WHERE.

FIG. 7A is an optimization of FIG. 7B by an algorithm for optimizing a subquery (Ref .: On Optimizing an SQL-like Nested Query, ACM Transaction on Database Systems, Vol 7, No. 3, September 1982, Pages 443-469). Can be converted to SQL. Conversely, this means that even if the <query> edge of FIG. 6b is converted to the <join> edge, the query graph has the same meaning. Therefore, even if all <query> edges in the query graph are replaced with <join> edges, the query becomes semantically equivalent. Therefore, in practice, it is much more convenient to generate the SQL statement of FIG. 7B from the query graph in which all the <query> edges are converted to the <join> edges than the SQL statement of FIG. 7A is converted to the SQL of FIG. 7B.

After the optimized SQL is generated as described above, the query executor 50 performs a query using the generated SQL to search for an object, and finally, the object-XML converter 60 finally retrieves the searched object from XML. Convert to (S40).

As described above, an apparatus and method for converting an ebXML filter query using G-FQL according to the present invention are a relational model, an object model, or a semi-structured process in the process of converting an ebXML filter query into a query language provided by a database management system. EbXML is based on the graph model base that can express queries on various data models such as models and converts the query query provided by the ebXML registry to SQL using G-FQL, a query word defined by XML syntax. When the filter query is modified and extended, it has the advantage of minimizing the update of the query processing device of the database management system. The ebXML filter query can be generalized by converting not only SQL but also various database query such as OQL or XQuery.

In addition, since ebXML is an international standard for electronic transactions, it is still being upgraded while accepting additional requirements for e-business between enterprises. Therefore, addition and modification of the ebXML filter query are also expected to be inevitable. An apparatus and method for converting an ebXML filter query using G-FQL according to the present invention have the advantage of minimizing the change of the query processing apparatus of the existing database management system and easily upgrading the apparatus.

What has been described above is only one embodiment for implementing the apparatus and method for converting ebXML filter quality using G-FQL according to the present invention, the present invention is not limited to the above-described embodiment, the scope of the claims Without departing from the gist of the present invention claimed in the present invention, anyone of ordinary skill in the art will have the technical spirit of the present invention to the extent that various modifications can be made.

Claims (5)

A G-FQL converter for converting a user-defined ebXML filter query to G-FQL, which is a graph model-based query word defined in XML syntax; An XML-object converter that converts G-FQL of XML syntax into a user-operable program object; A query graph generator for generating a query graph having a meaning equivalent to G-FQL from the G-FQL converted to the object; An SQL generator for generating SQL by optimizing a query in the query graph; A query processor that executes a query by using the optimized SQL generated; And The apparatus for converting ebXML filter quality using G-FQL, characterized in that the object-XML converter for converting the object retrieved by the query performer back to XML. The ebXML filter using G-FQL according to claim 1, wherein the G-FQL converter converts a user-defined ebXML filter query into a G-FQL expressed in XML syntax using a known XSLT engine and XSL. Query translator. The method of claim 2, wherein the basic elements of the query expressed in the G-FQL are a query element, a branch element, a subquery element and a filter element. EbXML filter quality converter using FQL. The method of claim 1, wherein the query graph generator refers to a vertex [Root] which means a starting point of the query graph and maintains a reference to the object to be queried, and means a query target object and A query consisting of four edges containing a plurality of vertices [Object] containing an identifier, a type, and an attribute, and a predicate representing the relationship between the object and the object, and the operation between the start and end object properties. An ebXML filter quality converter using G-FQL, characterized by converting into a graph. Converting the user-defined ebXML filter query to G-FQL, which is a graph model based query defined by XML syntax; Converting the G-FQL of the XML syntax into a program object operable by the user and generating a query graph having the same meaning as the G-FQL from the G-FQL converted into the object; Generating SQL by optimizing a query in the query graph; And Performing a query using the optimized generated SQL to retrieve the object and convert the retrieved object back to XML Method for converting ebXML filter quality using G-FQL, characterized in that consisting of.