JP2001236352A - Method and device for semistructured document retrieval and storage medium stored with semistructured document retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and a device for semistructured document retrieval which can retrieve document-in-XML information by an SQL enabling retrieval by obtaining a value from an XML document for an item specified by a conventional SQL in use at a retrieval request, dynamically generating a table, and performing conditional processing and a storage medium stored with a semistructured document retrieval program. SOLUTION: This invention has a retrieval language analyzing means which analyzes a retrieval language for retrieval from a database including a relational database represented in an input table form, a semistructured document analyzing means which analyzes document elements and the syntax of a semistructured document represented in the hierarchical structure of the document elements, a full retrieving means which obtain the values of document elements of the semistructured document meeting retrieval conditions analyzed by the retrieval language analyzing means, and a relational algebra operation means which outputs a retrieval result in table form by using the relation between the document elements according to the values of the document elements.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for retrieving a semi-structured document and a storage medium storing a program for retrieving a semi-structured document. The present invention relates to a semi-structured document search method and apparatus for searching for a document element and information in a semi-structured document expressed by a hierarchical structure thereof, and a storage medium storing a semi-structured document search program.

More specifically, by maintaining the relationship between elements in a document, a value is obtained from a document for an item specified by a search formula at the time of a search request, a table is dynamically created, and condition processing is performed. The present invention relates to a semi-structured document search method and apparatus using a search language that can execute a search by performing the search, and a storage medium storing a semi-structured document search program.

[0003]

2. Description of the Related Art In recent years, XML (eXtensible Markup Lang
uage: an extensible marked language) is attracting attention as an information distribution format, and is expected to be applied to e-commerce. Therefore, a technique for efficiently searching an XML document is required. An XML document is represented by a document element and its hierarchical structure. As search languages for XML documents, XQL and XML-QL have been proposed, but are not standardized. On the other hand, a database plays an important role in an existing system for managing and retrieving information, and in particular, a relational database is widely used.

[0004] SQL is widely used in the world as a standard for languages dealing with relational databases.

[0005]

However, the presence of both an XML document and a relational database as information sources causes the following problems.

(1) Inconsistency of search interface Relational databases in tabular form have become widespread.
There are many applications that use SQL as a search language. Such applications cannot search for information in an XML document represented by a hierarchical structure.

(2) Deterioration of search efficiency When desired information is distributed in a relational database and an XML document, it is necessary to search from each information source, and the efficiency is very low.

(3) Inconsistency of search results Since the format of search results from each information source is inconsistent, it is difficult to compare search results.

The present invention has been made in view of the above points, and obtains a value from an XML document for an item designated by SQL at the time of a search request using a conventional SQL, and dynamically creates a table. Provided are a semi-structured document search method and apparatus capable of searching for information in an XML document by SQL, which can be created and executed by performing conditional processing, and a storage medium storing a semi-structured document search program. The purpose is to:

[0010]

FIG. 1 is a diagram for explaining the principle of the present invention.

The present invention (claim 1) provides a semi-structured document search method for searching for a document element and information in a semi-structured document represented by a hierarchical structure of the document element. A relation between elements of a semi-structured document expressed by a hierarchical structure of elements is held (step 1), and a search expression of a search language for searching a database including a relational database expressed in a table format is provided. The value of the element of the semi-structured document is obtained using the element name of the semi-structured document for designating the item (step 2). Is performed (step 4), and arbitrary information in the semi-structured document is returned in one tabular form (step 5).

According to a second aspect of the present invention, there is provided a semi-structured document search method for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element. Analyzes a search language for searching a database including a relational database represented by, analyzes a document element and a syntax of a semi-structured document represented by a hierarchical structure of the document element, and analyzes a half-structured document corresponding to the search item. A grouping number is assigned to the analysis result of the structured document, and the position of the element corresponding to the item specified in the search language on the semi-structured document hierarchical structure is obtained based on the result of the analyzed semi-structured document. , Detecting a parent element name common to all items with grouping numbers, taking a direct product between groups, generating a result list, performing a relational algebra operation on the result list, It performs a comparison of the conditions for, and returns only the results that meet the condition as a search result.

According to the present invention (claim 3), JOIN of a semi-structured document is specified in a search expression of a search language, and information in the semi-structured document in a plurality of semi-structured documents is JOINed into a single table format. return.

According to the present invention (claim 4), a UNION of a semi-structured document is specified in a search expression of a search language, information in the semi-structured document in a plurality of semi-structured documents is unioned, and a single table format is used. return.

FIG. 2 is a block diagram showing the principle of the present invention.

The present invention (claim 5) is a semi-structured document search apparatus for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element. A search language analyzing means 110 for analyzing a search language for searching a database including a relational database expressed in a table format, and analyzing a syntax of a document element and a semi-structured document expressed by a hierarchical structure of the document element. Semi-structured document analysis means 120 and search language analysis means 11
All search means 140 that obtains the value of the document element of the semi-structured document that matches the search condition analyzed by 0, and outputs the search result in the form of a table using the relationship between the document elements based on the value of the document element And a relational algebraic calculation means 150.

The present invention (claim 6) further comprises a grouping means for assigning a grouping number to the analysis result of the semi-structured document corresponding to the search item analyzed by the search language analyzing means 110, The position of the element corresponding to the item specified by the search language on the semi-structured document hierarchical structure is acquired based on the result of the semi-structured document analyzed by the semi-structured document analysis means 120, and the grouping number Has a means for detecting a parent element name common to all items given with, taking a direct product between groups and generating a result list, and performing a relation algebra operation on the result list in the relation algebra operation means 150. And means for comparing conditions with the result list and returning only results that satisfy the conditions as search results.

According to the present invention (claim 7), the total search means 140
In the search expression of the search language, JOIN of the semi-structured document
And JOIN of information in the semi-structured document in a plurality of semi-structured documents.

According to the present invention (claim 8), the total search means 140
In the search formula of the search language, UNIO of the semi-structured document
Means for designating N and unioning information in the semi-structured document in a plurality of semi-structured documents.

The present invention (claim 9) is a storage medium storing a semi-structured document search program for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element. A search language analysis process for analyzing a search language for searching a database including a relational database expressed in an input table format;
A semi-structured document analysis process for analyzing the syntax of a semi-structured document represented by a document element and a hierarchical structure of the document element;
A search process that obtains the values of the document elements of semi-structured documents that match the search conditions analyzed by the search language analysis process, and the search results are tabulated using the relationships between the document elements based on the values of the document elements And a relational algebra operation process that outputs

The present invention (Claim 10) further comprises a grouping process for assigning a grouping number to the analysis result of the semi-structured document corresponding to the search item analyzed by the search language analysis process. Based on the result of the semi-structured document analyzed in the semi-structured document analysis process, the position in the semi-structured document hierarchical structure of the element corresponding to the item specified in the search language is acquired, and a grouping number is assigned. Detecting a parent element name common to all the items, taking a direct product between the groups, and generating a result list, performing a relational algebra operation on the result list in a relational algebra operation process, Has a process of comparing conditions with and returning only results that satisfy the conditions as search results. The present invention (Claim 1
1) includes a process in which, in the entire search process, JOIN of a semi-structured document is specified in a search expression of a search language, and information in the semi-structured document in a plurality of semi-structured documents is JOIN.

According to the present invention (claim 12), in the entire search process, a UNIO of a semi-structured document is used in a search expression of a search language.
N, and a process of unifying the information in the semi-structured documents in the plurality of semi-structured documents.

As described above, in the present invention, the relationship between elements in the XML document is held, and the items specified in the SQL statement correspond to the elements in the XML document, and are specified in the SQL at the time of the search request. For the items, values are obtained from the XML document, a table is dynamically created, and if the obtained tables require conditional processing, conditional processing is performed and search results in a unified format are output. It is possible to search for XML document information that returns information in the XML document in one table format.

Further, according to the present invention, for a plurality of XML documents specified by an SQL statement, the relationship between elements of the XML document is held, and the above-described processing is performed. It is possible to perform a condition process, perform a JOIN between a plurality of dynamically created tables, and return information in a plurality of XML documents in a single table format, and search for a plurality of XML document information. .

Further, according to the present invention, for a plurality of XML documents specified by an SQL statement, the relationship between the elements of the XML document is held, and the above-described means are performed. , Performing conditional processing, performing UNION between dynamically created tables, and returning information in a plurality of XML documents in a single table format, to search for a plurality of XML document information.

[0026]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, terms used in the following description will be explained.

SQL sentence: A language for handling relational databases, and handles XML documents with a similar grammar.

Element: <element name> Content of element </ element name>
The unit of the content of the XML document represented by: Item: Specified by the phrase of the SQL statement. SQ in the following description
The L sentence specifies the element name of the XML document. Multiple XML
Even if the same element name exists in a document, it can be uniquely specified by distinguishing it in advance by the file name.

Attribute: additional information attached to an element in an XML sentence.

One-to-one relationship: A relationship in which when a certain element exists in the same hierarchy of an XML document, a corresponding element always exists.

Grouping number: A number assigned to identify a group representing a relationship to a search item. When the grouping numbers of search items are the same, the relationship is a one-to-one relationship.

The present invention handles the XML document in the same manner as the information managed in the existing relational database, and constructs a system for efficiently retrieving information in a plurality of XML documents by SQL in the following points. To solve.

(1) Unification of search interface: By using SQL as a search language, it is possible to search information in an XML document from a conventional application.

(2) Search from a plurality of XML documents: A search method is used in which information in XML documents in a plurality of XML documents is JOIN or UNION and returned in one table format.

(3) Unification of search results: The search results from the XML document are tabulated to unify the search results from the existing relational database.

In order to perform the above, it is necessary to map the hierarchical structure model to a relational model. For this purpose, a method is considered in which an XML document is collectively divided according to a hierarchical structure before retrieval and stored in a relational database. However, the method of converting an XML document into a relational model collectively has a performance merit such as high-speed search, but a slight change in the logical structure of the document affects the database schema. It is not suitable for managing frequently changed documents. Therefore, there is a high possibility that the content and structure will be changed.
In order to construct a system for searching information in an L document by SQL, the following points need to be solved.

(4) Mapping to a dynamic relational model: For items specified in SQL at the time of a search request,
The value of the element is obtained from the XML document and dynamically mapped to a relational model.

FIG. 3 shows the configuration of a device for retrieving information in a semi-structured document using a database retrieval language according to the present invention.

The user inputs an SQL search sentence to the search device 100 via the application program 300, and the search device 100
The search device 100 searches for the information specified in the search sentence,
The search device 100 searches the XML document 200 for information specified in the SQL search sentence, and returns the search results to the application program 300 in a table format.

The search apparatus 100 shown in FIG. 1 includes an SQL sentence analysis module 110, an XML parser 120, a grouping module 130, an all search module 140, and a relational algebra operation processing module 150.

The SQL sentence analysis module 110 analyzes the SQL sentence input from the application program 300 and extracts a search result item, a search condition item, a relational algebra operation item, and a search target information source.

The XML parser 120 reads the XML document specified as the information source to be searched, parses the XML document, and expands the elements into a hierarchical structure.

The grouping module 130 gives a grouping number.

The all search processing module 140 extracts the value of the element of the XML document corresponding to the item specified as the search condition item.

The relational algebra operation processing module 150 performs processing on the extracted XML document values based on the search condition items and the relation algebra operation items, and obtains a search result.

The above-described information retrieval apparatus 100 for XML document information in SQL includes: (1) SQL sentence analysis processing; (2) XML document analysis processing; (3) grouping processing; (4) full search processing; and (5) relational algebra calculation processing; The following five steps are performed.

[0047]

Embodiments of the present invention will be described below with reference to the drawings.

The above-described (1) to (5) of the search device 100
Will be described using an XML document shown in FIG. 4 and an SQL sentence shown in FIG.

FIG. 6 is a flowchart of the search process according to the present invention.

Step 100) SQL sentence analysis processing: The SQL sentence shown in FIG. 5 is input to the SQL sentence analysis module 110 via an application program.
SQL statement is analyzed and SELECT indicating the search result item
Items, WHERE items indicating search condition items, and FROM items indicating search target information sources are extracted. The SQL sentence shown in FIG. 5 specifies information included under the “reciple-book” element of the XML document shown in FIG. 4 as a table to be searched, and the ing element has the element “name” and the element “name” under the search condition of “butter”. "In
g "and an element" amount "as a result.

Step 200) XML document analysis processing:
FROM of SQL statement by the XML parser 120
The XML document specified in the item is read and its syntax is analyzed. FIG. 7 shows a state in which the elements of the XML document of FIG. 4 are expanded into a hierarchical structure.

Step 300) Grouping process: The grouping module 130 assigns a grouping number to a search item. The one-to-one relationship is XML
Referenced in the document type definition of the document. In the case of the SQL sentence of FIG. 5, the document type definition is <! ELEMENT in-list (ing, amout) +>, and the element ing and the element amount of the search item have a one-to-one relationship. Therefore, the same grouping number is assigned.

If there is no document type definition, the XML parser 40
The one-to-one relationship is estimated from the parent-child relationship of the hierarchical structure of the XML document analyzed by 0. This means that in the hierarchical structure of FIG. 7, the element ing and the element "amount" of the search item of the same hierarchy always exist one-to-one. Table 1 shows an example in which a grouping number is assigned to the search item of the SQL sentence in FIG.
Shown in

[0054]

[Table 1] If there is no document type definition as shown in the XML document of FIG. 8, a grouping number is assigned to the search item using the parent-child relationship of the hierarchical structure of the XML document analyzed by the XML parser 120.

The specific procedure will be described below. A number is assigned to the element of the XML document corresponding to the search item specified by the SQL sentence in the order of appearance in the XML document. In the XML document of FIG. 8, 0 is added to the name element, 1 is added to the ing element, and 2 is added to the amount element. The XML document is parsed by the XML parser 120.
This is analyzed as shown in FIG. 7 in the XML document of FIG. In the case of the same hierarchy, the elements of the XML document are arranged in order from the left. If there is a one-to-one relationship, the numbers are set to the same number, and the ing element and the amount element that use this as the grouping number are in a repeated relationship on the same layer in the hierarchical structure of FIG.
Since it is known that there is always a one-to-one correspondence, the number of the amount element is set to 1 and used as a grouping number (the same as the grouping number when there is a document type definition shown in Table 1).

Step 400) Full search processing: This processing is performed by the above-described full search processing module 140. FIG.
9 is a flowchart of a full search process according to an embodiment of the present invention.

Based on the result analyzed by the XML parser 120, the SELECT item specified by the SQL statement and the W
The position of the element corresponding to the HERE item on the XML document hierarchical structure is obtained (step 401). In the case of the XML document of FIG.

[0058]

[Table 2] From the above acquisition results, a parent element name shared by all items is detected (step 402). In the case of Table 2, it is / recipe. Next, a common parent in the group is acquired (step 403). For group 0, it is / recipe, and for group 0, it is / ing-list.

Tracing from a common parent in the group, SQ
The process of acquiring the value of the element of the XML document corresponding to the item specified by the L statement is repeated by the number of items in the group. At this time, if there is an attribute, the attribute is acquired one-to-one with the element. The above process is repeated for the number of groups (step 404), and a direct product is obtained between the groups (step 40).
5). The above process is repeated for the number of common parents. XM in FIG.
Tables 3 and 4 show the results obtained by acquiring the values for the L document for the SQL in FIG. 5 and taking the direct product.

[0060]

[Table 3]

[0061]

[Table 4] Generate a result list from all records created by taking the direct product. Table 5 shows an example of the result list.

[0062]

[Table 5] Step 500) Relational Algebra Calculation Processing: The relation algebra calculation is performed on the result list (Table 5) obtained in the full search processing (step 400) using the above-described relation algebra calculation processing module 150. In the case of SQL in FIG. 5, WHE indicating a search condition item in the SQL analysis processing (step 100).
Since the RE items have been extracted, the conditions are compared with the result list (Table 5) obtained in the full search process (step 400). If the condition is "false", it is deleted and only the result that satisfies the condition is left. Table 6 shows the final results.

[0063]

[Table 6] Next, a specific example will be described in which a search is performed on a plurality of XML documents specified by an SQL sentence and the search is returned in one table format.

(1) In the case of JOIN: The SQL statement input through the application program 300 is
The QL sentence analysis module 110 analyzes. Figure 10 shows JO
This shows an SQL statement containing IN. F indicating the information source to be searched
A plurality of XML document names or the same XML document given an alias is specified in the ROM item.

The SQL sentence analysis module 110 decomposes the input SQL sentence into the number of XML documents specified in the FROM item. FIGS. 11 and 12 show a plurality of SQL sentences obtained by decomposing the SQL sentence of FIG.

The SQL sentence of FIG. 11 is the XML sentence of FIG.
Specify the information contained under the recipe-book element as a table to be searched.

The SQL sentence in FIG. 12 specifies information contained below the element of the XML document in FIG. 13 as a table to be searched.

Then, the information in the XML document is searched by SQL for each XML document as described in the above embodiment, and a table is dynamically created.

The above-described relational algebra operation processing module 150
Thus, JOIN is performed between a plurality of created tables, and information in a plurality of XML documents is returned in one table format. Table 7
Shows the final result.

[0070]

[Table 7] (2) In the case of UNION: The SQL sentence analysis module 110 analyzes an SQL sentence input via the application program 300. FIG. 14 shows an SQL sentence including UNION. The SQL sentence analysis module 110
Decompose a plurality of SQL statements connected by UNION. FIG.
FIG. 1 shows a plurality of SQL statements obtained by decomposing the SQL statements of FIG.
5, shown in FIG.

Then, the information in the XML document is searched by SQL for each SQL sentence described in the above embodiment, and a table is dynamically created.

The above-described relational algebra operation processing module 150
, A union is made between the created plurality of tables, and information in the plurality of XML documents is returned in one table format. Table 8 shows the final results.

[0073]

[Table 8] In the above embodiment, the description has been given based on the configuration of the search apparatus shown in FIG. 3. However, the SQL sentence analysis module 110, the XML parser 120, the grouping module 130, the full search processing module 140, the relational algebra shown in FIG. The processing module 150 is constructed as a program and stored in a disk device connected to a computer used as a search device, a floppy disk, a portable storage medium such as a CD-ROM, and the like. By installing, the present invention can be easily realized.

The present invention is not limited to the above embodiments, but can be variously modified and applied within the scope of the claims.

[0075]

As described above, according to the present invention, by using SQL as a search language, it is possible to search for information in an XML document from a conventional application. Multiple X
The information in the XML document in the ML document is stored in JOIN or UNI
Turns ON, and it is possible to perform a search for returning in one table format. Further, since the search result from the XML document is in the form of a table, it can be unified with the search result from the existing relational database.

The contents and structure can be changed by acquiring the value of the XML document element corresponding to the item specified by the SQL statement at the time of the search request, dynamically creating a table, and performing conditional processing. It is possible to search for information in an XML document with high reliability.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a principle configuration diagram of the present invention.

FIG. 3 is a configuration diagram of an apparatus for searching for information in a semi-structured document using a database search language according to the present invention.

FIG. 4 is an example of an XML document according to an embodiment of the present invention (part 1);
It is.

FIG. 5 is an example of an SQL statement according to an embodiment of the present invention.

FIG. 6 is a flowchart of a search process according to an embodiment of the present invention.

FIG. 7 is an example in which elements of an XML document according to an embodiment of the present invention are developed in a hierarchical structure.

FIG. 8 is an example of an XML document without a document type definition according to an embodiment of the present invention.

FIG. 9 is a flowchart of a full search process according to an embodiment of the present invention.

FIG. 10 illustrates an SQL statement including a JOIN according to an embodiment of the present invention.
It is an example of a statement.

11 is an example (No. 1) of an SQL sentence obtained by decomposing the SQL sentence of FIG. 10 according to one embodiment of the present invention.

FIG. 12 is an example (part 2) of an SQL sentence obtained by decomposing the SQL sentence of FIG. 10 according to an embodiment of the present invention.

FIG. 13 is an example (part 2) of an XML document according to an embodiment of the present invention.

FIG. 14 is an SQ including a UNION according to an embodiment of the present invention;
This is an example of an L sentence.

FIG. 15 is an example (part 1) of an SQL sentence obtained by decomposing the SQL sentence of FIG. 14 according to an embodiment of the present invention.

FIG. 16 is an example (part 2) of an SQL sentence obtained by decomposing the SQL sentence of FIG. 14 according to an embodiment of the present invention.

[Explanation of symbols]

REFERENCE SIGNS LIST 100 search device 110 search language analysis means, SQL sentence analysis module 120 semi-structured document analysis means, XML parser 130 grouping means, grouping module 140 full search means, full search processing module 150 relation algebra calculation means, relation algebra calculation processing module 200 XML document 300 application program

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Shiro Kasuga 2-3-1 Otemachi, Chiyoda-ku, Tokyo Within Nippon Telegraph and Telephone Corporation (72) Inventor Gengo Suzuki 2-3-1 Otemachi, Chiyoda-ku, Tokyo No. 1 F-term in Nippon Telegraph and Telephone Corporation (reference) 5B075 ND03 ND35 PP23 PQ02 PQ60 QS20 QT06 UU06

Claims (12)

[Claims] 1. A semi-structured document search method for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element, wherein the document element is expressed by a hierarchical structure of the document element and the document element. A search expression of a search language for searching a database including a relational database expressed in a tabular form by maintaining relationships between elements of a semi-structured document,
Using the element name of the semi-structured document for specifying the item, the value of the element of the semi-structured document is obtained, and if conditional processing is required, predetermined condition processing is performed. A semi-structured document search method characterized by returning arbitrary information in a structured document in one table format. 2. A semi-structured document search method for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element. Analyzing a search language for searching a database including the document, analyzing a syntax of a document element and a semi-structured document expressed by a hierarchical structure of the document element, and analyzing the semi-structured document corresponding to a search item A grouping number, and based on the result of the analyzed semi-structured document, obtain the position on the semi-structured document hierarchical structure of the element corresponding to the item specified in the search language, Detect a parent element name common to all the assigned items, take a direct product between groups, generate a result list, perform a relational algebra operation on the result list, And compares the condition, semi-structured document search method characterized by returning only the results that meet the condition as the search result. 3. A semi-structured document JOIN is specified in a search expression of the search language, and information in the semi-structured document in a plurality of semi-structured documents is JOIN.
3. The semi-structured document search method according to claim 2, wherein the search is performed in a single table format. 4. Specifying a union of a semi-structured document in the search expression of the search language, and converting information in the semi-structured document in a plurality of semi-structured documents to UN
3. The semi-structured document search method according to claim 2, wherein ION is performed and returned in one table format. 5. A semi-structured document search device for searching for a document element and information in a semi-structured document expressed by a hierarchical structure of the document element, the relational expression being expressed in an input table format. Search language analyzing means for analyzing a search language for searching a database including a database, semi-structured document analyzing means for analyzing the syntax of a document element and a semi-structured document represented by a hierarchical structure of the document element, A search unit that obtains a value of a document element of a semi-structured document that matches a search condition analyzed by the search language analysis unit; and a search result using a relationship between the document elements based on the value of the document element. And a relational algebra operation means for outputting a table in a table format. 6. A grouping unit for assigning a grouping number to an analysis result of the semi-structured document corresponding to the search item analyzed by the search language analyzing unit, wherein the full search unit includes the semi-structured unit. Based on the result of the semi-structured document analyzed by the document analysis means, the position on the semi-structured document hierarchical structure of the element corresponding to the item specified in the search language is obtained, and the grouping number is assigned. Means for detecting a parent element name common to all items, taking a direct product between groups, and generating a result list, wherein the relational algebra operation means performs a relational algebra operation on the result list, and 6. The semi-structured document search device according to claim 5, further comprising means for comparing a list with a condition and returning only a result satisfying the condition as a search result. 7. The whole search means includes means for designating a JOIN of a semi-structured document in a search expression of the search language and JOINing information in the semi-structured document in a plurality of semi-structured documents. A semi-structured document search device as described. 8. The semi-structured document according to claim 5, wherein said all-search means designates a UNION of a semi-structured document in a search expression of said search language, and UNIONs information in the semi-structured document in a plurality of semi-structured documents. A semi-structured document search device as described. 9. A storage medium storing a semi-structured document search program for retrieving a document element and information in a semi-structured document expressed by a hierarchical structure of the document element, the storage medium comprising: A search language analysis process for analyzing a search language for searching a database including a relational database represented by: and a semi-structure for analyzing the syntax of a document element and a semi-structured document represented by a hierarchical structure of the document element A structured document analysis process, a full search process for acquiring a value of a document element of a semi-structured document that matches a search condition analyzed in the search language analysis process, and a relationship between the document elements based on the value of the document element. And a relational algebra operation process for outputting a search result in a tabular format by using a storage medium storing a semi-structured document search program. 10. A grouping process for assigning a grouping number to an analysis result of the semi-structured document corresponding to a search item analyzed by the search language analysis process, wherein the full search process includes the semi-structured process. Based on the result of the semi-structured document analyzed in the document analysis process, the position on the semi-structured document hierarchical structure of the element corresponding to the item specified in the search language is obtained, and the grouping number is assigned. A process of detecting a parent element name common to all items, taking a direct product between groups, and generating a result list, wherein the relational algebra operation process performs a relational algebra operation on the result list, and 10. A process for comparing a list with a condition and returning only a result satisfying the condition as a search result.
A storage medium storing the described semi-structured document search program. 11. The all search process includes a process of designating a JOIN of a semi-structured document in a search expression of the search language and JOINing information in the semi-structured document in a plurality of semi-structured documents. A storage medium storing the described semi-structured document search program. 12. The whole search process includes a process of designating a UNION of a semi-structured document in a search expression of the search language and performing a UNION on information in the semi-structured document in a plurality of semi-structured documents. A storage medium storing the described semi-structured document search program.