KR100612376B1 - A index system and method for xml documents using node-range of integration path - Google Patents

Abstract

The present invention relates to an XML index system and method using a node range of an integrated path. After the integrated paths are integrated and configured, a number of preorders, post orders, node ranges, etc., are assigned to the integration path. The real data table names are stored in the dynamic array in the order of pre-order of the integration pass, so that the query is analyzed when the query is executed to obtain the preorder and node range in the integration path, and the data table name is obtained from the pre-order list dynamic array. Quickly analyze Xpath queries by efficiently opening them without comparisons, efficiently manage index storage space of paths, reduce the range of data tables to be actually searched using pre-order lists, and maintain data tables by lowest node. Easy insertion and deletion

XML, Indexing, Numbering, Node Range, Integration Pass, Preorder, Post Order

Description

System and method for XML index using node range of integrated path {A INDEX SYSTEM AND METHOD FOR XML DOCUMENTS USING NODE-RANGE OF INTEGRATION PATH}             

1 is a diagram illustrating an example of a structure of a general XML document in the form of a tree.

FIG. 2 illustrates the summarized simple path of FIG. 1. FIG.

3 shows an example of a data guide indexing technique.

4 shows an example of a Patricia Trie.

5 shows a Layered Index.

6 is an example of an index structure using an index fabric.

7 is an example of a numbering technique of Dietz.

8 is a diagram illustrating a conventional numbering-based SQL execution query example.

9 is an example implementation of the integration path of XML documents in the present invention.

10 is an index structure diagram when inserting an XML document in the present invention.

11 is an example of an existing numbering-based SQL query.

12 is a block diagram of an index system according to the present invention.

13A to 13E are explanatory views of data tables and fields in the present invention.

14 is a diagram illustrating an algorithm for performing a method for calculating a node range of a name index in the present invention.

15A and 15B are index structure diagrams before and after insertion of XML documents having different structures.

16 is a table showing experimental data in the present invention.

17a to 17c is a query, performance speed and performance comparison table and graph of the experimental data (A) in the present invention.

18a to 18c is a query, performance rate and performance comparison table and graph of the experimental data (B) in the present invention.

19a to 19c is a query, performance speed and performance comparison table and graph of the experimental data (C) in the present invention.

20 is a graph comparing the size of the index according to the present invention and XISS.

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

100: XML database 200: DOM parser

300: low data storage unit

The present invention relates to an eXtensible Markup Language (XML) indexing system and method, which integrates and manages all XML document paths and improves search performance by adding a Node Range and a Pre-Order List. The present invention relates to an XML index system and method using a node range of an integrated path.

Recently, XML is recognized as a standard for the representation and exchange of data on the Internet, and the importance of research is gradually increasing. There is an increasing demand for storing, retrieving, and exchanging important information on the Internet in XML. In addition to specialized fields such as medicine, defense, law, and tax, XML is used in almost all fields on the Internet such as e-commerce and messenger. It is used. In particular, specialized XML documents have become large in size and complicated in user queries. Therefore, many indexing techniques have been proposed to store numerous XML documents and to process complex XML queries quickly.

Among the typical indexing techniques, there are a pass-based indexing technique, an indexing technique based on a numbering technique, and many other indexing techniques. The indexing technique based on the double pass and the indexing technique based on the numbering technique, which has been the foundation technique in many studies, are discussed.

One). Pass-Based Indexing Technique

The XML document can be expressed in the form of a tree structure as shown in FIG. 1 and includes many duplicated nodes due to the characteristics of the document. There is a problem in that all paths including duplicate nodes must be operated in order to retrieve a corresponding XML document or data in the XML document. To solve this problem, Data Guide (hereinafter referred to as DataGuide) (R. Goldman, J. Widom. "DataGuides: enabling query formulation and optimization in semistructured databases", In Proc. Of the VLDB, pp. 436-445 , 1997.) has been proposed.

The XML document of FIG. 1 expresses nodes such as "Restaurant" and "Entree" in a duplicated form, and (1) to (11) represent locations of data stored in an actual database. If the user query "/ Restaurant / Name" is given, the XML document has the "Restaurant" tag as the root node, and all data including the "Name" tag as the child node must be retrieved. In order to solve this problem, the proposed DataGuide defines possible forms of all queries from the root as a simple path as summarized in the table of FIG. 2.

In the DataGuide indexing method as shown in FIG. 3, each node may be connected to a set of stored data. When executing query "/ Restaurant / Name", DataGuide retrieves Restaurant and Name from node in turn and returns data storage location of (5) and (9). To integrate the redundant paths proposed by DataGuide, 1-Index / 2-Index / T-Index (T. Milo, D. Suciu, "Index Structures for Path Expressions", In Proc. Of the ICDT, pp. 277- 295, 1997.) and Index Fabric (B. Cooper, N. Sample, M. Franklin, G. Hjaltason, M. shadmon, "A Fast Index For Semistructured Data", In Proc. Of the VLDB, pp. 341-350 , 2001.) has been proposed. In particular, Index Fabric uses Patricia Trie (D. Knuth, "The Art of Computer Programming, Vol. III, Sorting and Searching", Third Edition.Addison Wesley, Reading, which encodes and searches strings to reduce the pathing of complex XML documents. , MA, 1998.).

Patricia Trie of FIG. 4 is implemented by an encoding method of removing duplicate characters and storing positions of different characters and corresponding characters when inserting a string.

 The Index Fabric focused on reducing the range of path operations by defining the path and data from the root to the leaf of the XML document as raw data and encoding it as Patricia Trie. In addition, in order to form a balance tree, the management is divided into layers as shown in FIG. 5.

All paths are encoded using Designators. In addition, all nodes are managed as summarized characters through digitizers and mappings in the Designators Dictionary. Summarized characters consist of summarized simplepaths and raw data with the bottommost XML data listed. The index is constructed using Patricia Trie as shown in FIG. 6 using the raw data. Index Fabric shows very good performance when searching the encoded path from the root node to the leaf node as shown in FIG. However, the efficiency of the search for the intermediate node, the lowest node, and the join operation of the ancestor-progenitor relationship ('//') is reduced. To compensate for this, we used the method of adding a frequently used path to the index by using the concept of refined-path, but it is not suitable to handle various queries because the DBA needs to define it manually, and in many cases, the effect of encoding is inferior.

2). Numbering-based indexing technique

In XPath query, the join operation ('//') of ancestor-descent relationship occurs frequently. One of the underlying methods for handling these operations is Dietz's Numbering Scheme (P. Dietz, "Maintaining order in a linked list.", In Proc. Of the Fourteenth Annual ACM Symposium on Theory of Computing, pp. 122-127, 1982.), and many recent indexing techniques use this method for ancestor-progenitor join operations.

Every node in an XML document based on the numbering technique has a number <d ₁ , d ₂ , d ₃ >. If nodes m and n exist and have a relationship of mn, if m is a descendant of n, then nd ₁ <md ₁ and nd ₂ > md ₂ , and m is an ancestor of n If nd ₁ > md ₁ and nd ₂ <md ₂ , then m is a child of n, then nd ₃ + 1 = md ₃ . In addition, if m and n are not ancestor- descendants, nd ₂ <md ₁ or nd ₁ > md ₂ is satisfied. That is, <d ₁ , d ₂ , and d ₃ > indicate <preorder, postorder, level>.

For example, if node a and node b of FIG. 7 are represented by the above numbering technique, node a has <1,7,1> and node b has <2,4,2>. Since node b is a descendant of node a, it satisfies 1 <2 and 7> 4, and node a satisfies 2> 1 and 4 <7 because node a is an ancestor of node b, and node b is a child of node a. .level + 1 = satisfy the relationship of node b.level Using the above structure, the index is constructed by assigning the numbers <d ₁ , d ₂ , d ₃ > to all nodes of the XML document, and performs user query processing. However, there is a problem in that the labels of all the nodes must be adjusted when the intermediate node or the child node is inserted. XISS (P. Harding, Q. Li, B. Moon, "XISS / R: XML Indexing and Storage System Using RDBMS", In Proc. Of the VLDB, pp. 1073-1076, 2003.), (Q. Li , B. Moon, "Indexing and Querying XML Data for Regular path expression", In Proc. Of VLDB, pp. 361-370, 2001.), extended the numbering technique to solve the above problem.

In XISS, all nodes are managed by extending them to order and size instead of preorder and postorder, and the size includes the number of data to be included in the child node + the expandable range. That is, if there are nodes m and n and m is a parent node of n, the following relationship is satisfied. If nodes m and n exist and n is a descendant of m, then order (m) <order (n) and order (n) + size (n) <= order (m) + size (m) You will be satisfied. Based on the above method, XISS manages element index, attribute index, and data index for data in every XML document.

<Order, Size> is given for each XML document, so it can be easily indexed even if a document with a different structure is inserted. However, the space of the index is inefficient because the document ID, the order and size of each node in the document must be managed for each document. When searching, join operations must be performed for each document ID, and as many self-join operations are performed as the number of nodes in the query, the search overhead is very large. That is, when there is a query such as "/ Book / Authors // Name", the numbering-based indexing techniques are performed as shown in FIG.

Since indexing is done for each XML document, join operations are required for each document ID (doc_id) as many nodes as included in the XPath query. Data is also compared with the number assigned to each document node. The larger the data and the larger the number of nodes included in the query, the lower the retrieval performance, and the worse the performance if the data or attribute values are compared.

As such, all existing methods based on numbering techniques manage each XML document by dividing each XML document into an index such as path (or element), data, and attributes. The path index assigns a unique document ID and a number for each node, and the data and attribute indexes sequentially assign a value between the document ID and the number for each node. Then, the user's query is analyzed and the search result is extracted through the join operation of the document ID and the number of each index such as path, data, and attribute. Therefore, the larger the number of XML documents and the larger the size of data, the larger the size of the path index. The larger the number of nodes included in the query, the higher the self-join operation of the path index.

The present invention has been proposed to solve the above-mentioned problems of the prior art, and an object of the present invention is to integrate and manage a path of an XML document, and to reduce the size of an index because data and attribute indexes in the document are not given a number. In addition, to provide an XML index system and method using the node range of the integration path with excellent search performance.

According to an aspect of the present invention, an XML index system using a node range of an integrated path includes: a DOM parser configured to input XML documents into simple paths and data; A row data storage unit for temporarily storing the output of the DOM parser; The integrated path is configured with respect to the data from the raw data storage unit, compared with the existing path, and the pre-order, the post-order, the node range indicating the range of the corresponding data, etc. are compared with the previously stored information about the data. Name index for storing; A pre-order list for organizing all nodes of the integration path in pre-order order to store the name of the data table; An XML document ID table for storing an ID of an XML document and an XML document name for each name of the data table; And an XML database including an ID of the XML document, a data table in the document, and a data table in which the order of the data values is stored.

In order to achieve the above object, the XML indexing method using the node range of the integration path according to the present invention comprises a plurality of XML documents in one path through the same node to configure the integration path, but is configured by the DOM parser It is characterized by managing a separate index constituting XML data for each document ID for each node of the simple path.

In addition, a node range indicating a preorder, a postorder, and a corresponding data range is given to each node configured in the same path, and all nodes configured in the same path are arranged in a preorder order, and a real data table in the preorder order. It is to be stored in the name of the user query, the query is analyzed to obtain a preorder and node range in the integration path, and get the corresponding data table name stored in the order of the preorder, characterized in that the open without a number comparison.

Hereinafter, the present invention will be described in more detail with reference to the accompanying drawings. However, the following examples are merely to illustrate the present invention is not limited to the contents of the present invention.

The present invention integrates and configures XML documents having various structures as one path through the same node as shown in FIG. 9. In addition, it manages separate indexes that constitute XML data for each document ID by the lowest node of a simple path such as "Book / Titles / Title", "Book / Authors / Author / Name", and "Book / Store / Name". In other words, only the number of paths is assigned and managed, and the path search result is used to bring only the range of the table where the data is stored.

When all XML documents are inserted, the index structure according to the present invention is shown in FIG. Each node in the consolidated XML pastry is given a node range representing the preorder, postorder, and range of data using a numbering technique. The Pre-Order List configures the nodes of the integration path in a preorder order, and each item stores the name of the data table that stores the data of the lowest node.

When a user's XPath query is given, the integration path is used to find only the range of data and immediately open the corresponding data table.

For example, when an XPath query of "Book // Author" is given to the index system having the structure of FIG. 10, the method is implemented based on the existing numbering method.

A data index (value_ix) in which data and attributes are stored in the range obtained through the join operation of the document ID (doc_id) of the element index (Element _ix) and the number (order, size) of each document node. ). The larger the number of nodes included in the query, the greater the search overhead due to more self-joins.

Therefore, the present invention integrates and manages all XML document paths, so the path size is very small. Therefore, it is very efficient to compare and search the number given to the node in XPath query. That is, when performing a query of "/ Book // Author" in FIG. 11, the present invention results in <1,15,15> and <7,8,10> of "<Book>" and "<Author>". If we obtain and satisfy the ancestor-descendant relationship between preorder and postorder, we get <preorder, node range>, that is, <7,10>. It directly moves from the 7th block to the 10th block of the pre-order list and returns the data table name in the block. By directly opening the returned data table itself, you get a query result. The search performance is excellent because the number comparison in the data index is not performed as in the conventional numbering technique.

The configuration of the index system according to the present invention is as shown in FIG.

First, as shown in FIG. 12B, the XML database 100 of the present invention constructs an integration path, and includes a name index 110 for analyzing an XPath query, and preorders all nodes of the integration path. A pre-order list 120 for storing the name of the data table, an XML document ID table (Doc_id_TB) 130 for storing IDs of XML documents included in the data table, and actual data. It consists of a data table (Data Table) 140 to store, the details of which are as shown in Figs. 13A to 13E.

In addition, the index system of the present invention configures a simple path of all data using the DOM parser 200 when inserting an XML document as shown in FIG. 12A. For example, in the case of the XML document of FIG.

/ Book / Titles / Title: C ++

/ Book / Store / Name: Young, Kim

The simplepath and the data of the data, the raw data of the simplepath and the data in this way is temporarily stored in the row data storage unit 300, the name index 110 for configuring the information of the XML data in the stored data It compares with the information saved in the previous and saves by giving <preorder, postorder, node-range>. In other words, rather than assigning a number in XML document units, it is managed as an integrated path compared to the existing path. The node range of the name index 110 is used to quickly move to the pre-order list 120 using the range of data obtained by analyzing the XPath query.

14 shows a method of calculating a node range.

There are two main ways to calculate the node range. If the current node is the root node, the scope of the current node is the root node's postorder. If it is not a root node, if there is a right-sibling node, the value of the preorder-1 of the right sibling node is the node range of the current node, and if the right sibling node does not exist, the parent node The node range of becomes the node range of the current node.

The pre-order list 120 of FIG. 13B is one dynamic array stored in the preorder order of the name index 110 of FIG. 13A. Each item stores the name of the preorder of the name index 110 and the data table. It is very small because it only stores the names of the data tables in the preorder order of the combined pass.

The name of the data table is automatically given a unique name for each lowest node added when an XML document is inserted. When processing an XPath query, the lowest preorder and node range are obtained through the name index 110. The obtained preorder and node range are used to move directly to the location of the pre-order list 120 and to obtain the name of the data table.

The XML document ID table 130 stores the unique number of the XML document and the name of the XML document for each data table name of the pre-order list 120. It is used to return the result of a query as an archived XML document.

The data table 140 is generated for each data table name stored in the pre-order list 120, and the document unique number (Doc_id) of the XML document, the data value (Xml_value) and the order of the data value (Seq_value) in the document are Will be saved. The order of the data values is given a sequence number if they have the same tag (node) name in the XML document. It is used to handle "/ Book / Authors / Name" of XPath query. Since the data table 140 is generated for each of the lowest nodes, when the result range <preorder, node range> of the XPath query is wide, or when a large number of data tables are opened due to the presence of duplicate nodes, a lot of search-overheads are generated in the XML database 100. In this case, the data table name may be designated as a cluster index in the single data table and processed as shown in the data table 150 of FIG. 13E.

Next, we will look at the insertion of XML documents.

The present invention has a separate data table for each lowest node and operates as an integrated path. Therefore, if an XML document having the same structure is inserted, no path adjustment is necessary. The process is completed by adding data to the data table registered in the corresponding pre-order list 120 along the path. The data table does not have a number, so there is no processing.

If an XML document having a different structure is inserted, the name index 110 and the pre-order list 120 need to be adjusted. That is, the node is added to the name index 110 and the number of the changed portion is recalculated, but the overhead is not large because the size of the name index 110 is small. The number of arrays in the pre-order list 120 is increased, but the existing table names for each node are not changed. However, the insert process is completed by automatically assigning only the table name corresponding to the added node and storing the data in the corresponding data table.

15A and 15B illustrate an example of inserting XML documents having different structures. FIG. 15A shows an index structure before insertion and FIG. 15B shows an index structure after insertion.

Even when inserting an XML document having a different structure as shown in FIG. 15B, <preorder, postorder, noderange> of the corresponding position of the name index 110 is adjusted, and only the name of the data table added to the pre-order list 120 is automatically. The processing is completed by granting.

By comparing the present invention with the existing indexing technique, it is intended to prove that the present invention contributes to the improvement of index space and search performance.

First, since the index to be compared is very inefficient in intermediate nodes and ancestor- descendant join operations, the index is compared and evaluated with XISS, which is one of the typical indexing techniques based on the existing numbering technique. Implement and process dynamic arrays, but there may be memory limitations when many XML documents with different structures are inserted. Therefore, build and compare on disk.

The index system of the present invention is implemented on a PC with a Pentium 4 2.0G CPU, 512MB RAM, and 60GB HDD, using the MS-Windows 2000 server as an operating system, and the relational database used for the implementation is MS-Sql * Server 2000. For development language, Visual Basic 6.0 and T-SQL were used. In addition, Microsoft's MSXML (DOM) library was used as a parser to extract the internal data and structure of experimental XML documents. The experimental data were largely divided into three categories (A), (B), and (C). Experimental data (A) is composed of legal XML documents currently operating on the actual legal site. Each document has a very large number of documents, although the number of nodes and data contained in the document is small. Experimental data (B) is XISS-R (R. Goldman, J. Widom. "DataGuides: enabling query formulation and optimization in semistructured databases", In Proc. Of the VLDB, pp. 436-445, 1997.). In Shakespeare's comedy XML documents used as experimental data in, the number of documents is small, but the size of each document is very large because the number of nodes and data included in each document is very large. Experimental data (C) is an arbitrary synthetic dataset that consists of a large amount of XML documents that contain many duplicate nodes and data and have very deep features.

Experimental data (A), (B), and (C) of different structures were used to compare the index system of the present invention with the existing XISS / XISS-R system. We compared and analyzed the size of index data.

16 is experimental data, (A) is an actual XML document of a legal site, (B) is a XISS-R (Shakespeare play), and (C) is a synthetic dataset.

Each experimental data (A), (B), (C) shows a large difference in the number of average elements per document, the number of average data, and the depth of the document.

The experiment data (A) analyzed a search time using a total of 12 queries, the experimental data (B) and the experimental data (C) a total of nine queries. There are three types of queries: simple path queries, ancestor- descendant relational queries, and ancestor- descendant conditional queries. The retrieval time is the total of the pre-order (1), the pre-order (2), and the pre-order list using XISS and the method proposed by the present invention in Figs. 13A and 13E based on disk rather than dynamic array. Four indexing methods were used to compare. In addition, the size of the index for each experimental data was also compared and analyzed.

The experimental data (A) is characterized by having a very large number of documents but a small number of nodes and data contained in each XML document. There are a total of 12 queries used in the experimental data A, as shown in FIG. 17A. Q1 through Q3 consist of a simple pass or a simple pass from a root node to a leaf node, and Q4 through Q6 form a join operation of an ancestor-descendant relationship, and Q7 represents a complex structure of an ancestor-descendant relationship and a simple path. . Q8 to Q12 are constructed in a condition that contains the query structure from Q1 to Q7.

FIG. 17B is a table comparing XISS and the method according to the present invention as a result of performing the 12 queries of FIG. 17A and a method of implementing the data table and the pre-order list of FIG. 13D and FIG. 13E on disk. Indicates the search time and the number of result records returned. 17C graphically depicts the results. It can be seen that the search results are superior to the existing methods in all queries. The size of the integration path itself is also smaller than the XISS, which is indexed for each node in each document, and the data is also searchable because the results are returned by opening the data table in the pre-order list without comparing the numbers from Q1 to Q7. . In addition, in case of XISS, there were many differences in execution time according to the amount of data. In case of condition query from Q8 to Q12, not only nodes but also data should be searched. However, the existing search method reduced the search performance, but this method showed excellent search performance. The disk-based implementation of the pre-order list also performed better than the conventional method.

Although experimental data (B) has a small number of XML documents, the number of nodes and data for each XML document is very large, and the structure is also complicated. As in the experimental data (A), the simple path or the simple path was omitted from Q1 to Q3, the join operation of the ancestor-progenitor relationship from Q4 to Q7, and Q8 and Q9 were configured in the form of conditional queries (FIG. 18A).

As shown in the experimental data (A), the search performance was not significantly different. Both methods showed excellent search speed. However, in the case of conditional query Q8 which contains a lot of data, the proposed method showed excellent search performance. In addition, in the case of the conditional query in the experimental data (A), the method using the data table (cluster index) of FIG. 13E showed a little better performance, but the experimental data (B) showed a lower performance as the result of Q9. . That is, when the number of nodes and the amount of data in the document are very large, it can be seen that the performance of the data table of FIG. 13D is excellent. The disk-based method of the pre-order list showed excellent performance on average as in the case of the experimental data A (Figs. 18B and 18C).

Experimental data (C) is a synthetic dataset that has a deep level, contains many duplicate nodes and data, and has a large structural feature. Since Q1 and Q2 have a deep level, queries with simple structures omitted, Q3 through Q7 join-join queries of ancestor- descendant relationships, and Q8 and Q9 consist of ancestor-Descendant relationship + conditional queries (FIG. 19A).

Because of the characteristics of the experimental data (C), the sample XPath query is composed of a query that includes many nodes because it includes many nodes in the document and the level of the document is deep. In the case of XISS, the larger the number of nodes included in the query, the more self-join operations occur, so that the search performance is lower than that of the present invention. In addition, in the case of Q8 and Q9, the performance is lowered because the number of data indexes is compared and the condition is searched. In this method, even though the number of nodes included in the query is high, the average performance is excellent, and the conditional query of Q8 and Q9 has more than doubled the performance. As a result of comparative experiments using the experimental data of different structures (A), (B), and (C), the proposed method showed excellent search performance on average regardless of the data structure. In addition, when a large number of documents with different structures are inserted, the pre-order list becomes very large, and the performance is evaluated by implementing on a disk basis instead of a dynamic array. As a result, the performance is lower than that of the dynamic array, but the search performance is superior to the conventional method. As a result, when the size of the stored XML documents is very large, the search performance is excellent, but the size of the document is small and the search performance is more excellent in many cases. In addition, while the existing method depends on the number of nodes included in the query, the search performance is not significantly affected in the present method (Figs. 19B and 19C).

20 is a graph comparing the size of the index according to the present invention and XISS. Since the present invention integrates and manages paths and does not assign a number to each data table, it can be seen that the index space is managed more efficiently than the conventional method.

As described above, although described with reference to a preferred embodiment of the present invention, those skilled in the art various modifications of the present invention without departing from the spirit and scope of the invention described in the claims below Or it can be changed.

As described above, the XML index system and method using the node range of the integration path according to the present invention is configured by integrating the overlapping path, give a number of preorders, post orders, node range, etc. to the integration path, The name of the real data table is stored in the dynamic array called the order list in the order of pre-order of the integration path, so that the query is analyzed when the query is executed to obtain the preorder and node range in the integration path and the corresponding data table in the pre-order list dynamic array. Quickly analyze Xpath queries, efficiently manage index storage space of paths, reduce the range of data tables to be actually searched using pre-order lists, and place data tables at the lowest level by getting names and opening them immediately without number comparison. It is easy to insert and delete by keeping by node.

Claims (11)

In the XML index system, A DOM parser configured to input XML documents into simple paths and data; A row data storage unit for temporarily storing the output of the DOM parser; Compose an integrated path with respect to the data from the raw data storage unit, compared to an existing path, and give a preorder, a post order, a node range indicating a range of the corresponding data, and the like by comparing the previously stored information with respect to the data. A name index to store; A pre-order list for organizing all nodes of the integration path in pre-order order to store the name of the data table; An XML document ID table for storing an ID of an XML document and an XML document name for each name of the data table; And an XML database comprising an ID of an XML document, a data table storing data values and a sequence of data values in the document. The XML using the node range of the integrated path according to claim 1, wherein the name of the data table stored in the pre-order list is automatically given a unique name for each lowest node added when the XML document is inserted. Index system. The method of claim 1, wherein the data table XML index system using a node range of the integration path, characterized in that it is generated for each name of the data table stored in the pre-order list. The method of claim 1, wherein the order of data values stored in the data table is as follows. XML index system using a node range of the integration path, characterized in that the order is given when the same tag (node) name in the XML document. The XML index using the node range of the integrated path according to claim 1, wherein the data need only be added to a data table registered in a corresponding pre-order list along a path when the XML document having the same structure is inserted. system. The method of claim 1, wherein a node is added to the name index when the XML document having a different structure is inserted, the data table name corresponding to the added node is automatically assigned, and the data is stored in the corresponding data table to complete the insertion process. XML index system using the node range of the integrated path characterized in that. Integrate multiple XML documents into one path through the same node to form an integration path, but manage separate indexes that constitute XML data for each document ID for the lowest node of the simple path configured by the DOM parser. XML indexing method using the node range of the integrated path. 8. The XML indexing method according to claim 7, wherein a node range indicating a range of preorders, post orders, and corresponding data is assigned to each node configured in the same path. 10. The method of claim 8, wherein all nodes configured in the same path are arranged in a preorder order. 10. The method of claim 9, wherein the names of the real data tables are stored in the order of the preorders. The node of an integrated path according to claim 10, wherein the user query is analyzed to obtain a preorder and node range in the integration path, obtain a corresponding data table name stored in the preorder order, and immediately open the number without comparison. XML indexing method with scope.

Images