JPH096803A - Document data base management device - Google Patents

Abstract

PURPOSE: To make it easy to judge the adequacy of a retrieval expression by collating a document type with the conditions that are specified with the retrieval expression and use parent-child and descendant relation. CONSTITUTION: A correspondence table holding part 9 holds a correspondence table which has as entries three sets of elements (starting element), elements (adjacent element set) appearing as children of the starting element, and a set (reachable element set) of elements that can be reached from the starting elements. Further, a correspondence table generation part 10 refers to the structure of a specified document type as to the document type and generates the correspondence table. A reaching possibility decision part 11 when given two elements which are a source and a destination scans the correspondence table held at the correspondence table holding part 9 to inspect whether or not the destination is included in the adjacent element set or reachable element set of an entry based upon an entry as a starting element. When a structured document is retrieved, collation is performed under conditions which are specified with the retrieval expression and use the parent-child and descendant relation to judge whether or not the retrieval expression is adequate.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document database management device for managing electronic documents.

[0002]

2. Description of the Related Art Since an electronic document created by a word processor or the like is expressed as digital data, it is possible to easily add, delete, change, etc., and improve the efficiency of document creation. Further, by building a document database device by accumulating a plurality of electronic documents in a large-capacity storage device, it is possible to electronically search for a target document by keyword search or the like.

In a conventional document database management apparatus that manages electronic documents, when searching for documents,
The document data itself created by a word processor was stored, and the data was used for searching.

On the other hand, in order to facilitate the creation and editing of electronic documents, the electronic documents are structured. The structure of a document is represented by information about the elements such as chapters, headings, and paragraphs that make up the document and the relationship between the elements, for example, a chapter has headings and paragraphs as substructures. It

[0005]

The problem to be solved by the present invention is solved by the ODA (Off) which is an international standard for document structure.
ice Document Architecture
e) (ISO 8613) and SGML (Standard)
d Generalized Markup Lang
age) (ISO 8879; JIS X415
1) will be described as an example.

First, terms used in this specification will be described.

The term "document structure" is an information structure that represents a document. For example, the information structure defined by ODA is a document structure. The document structure is also one in which SGML subsetting (function limitation) is performed and the information structure used for the character codes and charts used is defined. Note that S
For GML, for example, Martin Brya
n, "Introduction to SGML", ASCII Corporation, 1991
Please refer to the publication on March 31, 2013.

The term "document type" shall mean a template for a document. The document type defines what kind of logical structure a document created from can have, that is, the types of nodes that appear in the logical structure, the attributes that each node can have, and the substructure that each node can have. ODA common logical structure (generic logical str)
structure), and DTD (Document) in the document architecture in which SGML is set.
Type Definition) is a document type.

Next, a problem in the case of retrieving a structured document as described above will be described.

In a structured document, the content of the document is called a logical structure, and is represented by a tree structure composed of a plurality of document constituent elements such as chapters, sections and figures. An example of the logical structure is shown in FIG.

The logical structure is not completely free to be created, but is created according to the syntax rule called the document type described above. An example of the document type is shown in FIG. The rectangular node defines the element type (element type). The label of the node indicates the name of the element type. Entities of nodes having the same name have the same element type. Therefore, the element type named "section" in FIG. 11 is recursively defined. Nodes indicated by ellipses define the connection of elements. This node is called a constructor. SEQ nodes are:
This indicates that the instances of the nodes connected to it are generated in that order. The REP node indicates that an instance of the node connected to the REP node is generated once or more. The OPT node indicates that the instance of the node connected to it may or may not appear. CHO
The node indicates that an instance of any one of the nodes connected to the node is generated. The logical structure of FIG. 10 satisfies the document type constraint of FIG.

A document database management device that manages structured documents provides a query language for describing a search. Some query languages are described in text, and some are described in a graphical user interface. FIG. 12 shows an example of the search formula described by the graphical user interface. The character string of the node indicates the element type. The character string shown beside the node indicates that the text held by the node includes the character string. The arc indicated by the solid line indicates that the nodes at both ends have a parent-child relationship. The arcs indicated by broken lines indicate that the nodes at both ends have an ancestor relationship (relationship between ancestors and descendants). When multiple arcs are output from one node, the search results are those that satisfy all the conditions. In other words, it has been designated as a conjunction. The search formula of FIG. 12 specifies a search for "a chapter that includes a character string" document "in a heading and a paragraph that includes a character string" database "".

The search expression is specified by using the conditions regarding the elements of the document and the conditions regarding the connection relation between the elements. FIG.
In the example of 2, the former is a condition related to element types, and the latter is a condition using an ancestor relationship.

FIG. 13 is a schematic diagram of a conventional database management device for performing the above-described search. The search formula created by the query editor 1 is
The search expression is converted into a search expression that can be executed by the search expression evaluation unit 3. This search formula is passed to the document type management unit 5 to search for a document having an element that satisfies a specified condition. In addition, 6 is a data dictionary and 7 is a database. The search expression evaluation unit 3 also checks whether or not the search expression is grammatically correct. If the search expression is not grammatically correct, the fact is notified to the operator and the process is stopped.

The conventional document database management apparatus for managing electronic documents shown in FIG. 13 only checks whether or not the search expression is grammatically correct. For this reason,
Even if a search formula for which a solution cannot exist is given, it is treated as a correct search formula. For example, "in paragraph" as shown in Figure 14
The search expression "including the character string" database and including the character string "document" in the heading connected to the lower part of the paragraph "is grammatically correct, but no document has an element that satisfies this. That is, in the case of the structure of the document example shown in FIG. 11, it is impossible that a headline exists in the lower part of a paragraph.

As a result of searching such an invalid search expression, there is no one satisfying the condition, so that nothing is obtained. From the user's point of view, the search expression is grammatically and semantically correct, but it is not easy to determine whether or not the conditions that satisfy the conditions did not exist in the database, or the search expression was not valid in the first place. It was a burden on the user when constructing the search formula. Moreover, since the retrieval processing is performed even if there is no document having an element satisfying the condition, the calculation time of the system is meaninglessly wasted. The evaluation of the search expression shown in FIG. 14 typically scans all heading and paragraph instances,
After that, it is checked whether or not there is one that satisfies the parent-child relationship, but there is no one that satisfies the condition of FIG. 14 in the first place, so there is no need to evaluate.

Therefore, an object of the present invention is to make it possible to easily judge the validity of a search expression. Another object of the present invention is to prevent the search time from being wasted by not performing a search for an invalid search expression.

[0018]

In order to solve the above-mentioned problems, the present invention manages a structured document that defines a structure that a document can take according to a syntax rule, and sets a condition regarding an element of the document and an inter-element condition. Using the condition regarding the connection relation,
The document database management device for designating a search target is characterized by including a means for matching a syntax rule with a given search expression.

Further, according to the present invention, based on a structure that a document can take, a correspondence table is provided in which a starting element, an element subordinate to the starting element, and an element subordinate to the starting element are one set. And means for verifying whether the search expression is valid by scanning the correspondence table based on the search expression.

[0020]

According to the present invention, when retrieving a structured document, the document type and the conditions using the parent-child relationship and the grandchildren relation specified in the retrieval formula are collated to judge whether the retrieval formula is valid or not. To be done.

In the present invention, the document type is checked before executing the search, and a starting element, an element subordinate to the starting element, and a set of elements which can be subordinate to the starting element are grouped. A corresponding table is generated. When a search expression is input, the correspondence table is scanned based on this search expression, it is determined whether or not the search expression satisfies the conditions of the correspondence table, and when the conditions are not satisfied, the search expression is not valid. To be judged.

[0022]

1 is a block diagram of a document database management apparatus according to the present invention. The query editor 1 is for inputting search conditions and is input using a query editor screen as shown in FIG. The search formula created by the query editor 1 is processed by the search formula generation unit 2 and the search formula evaluation unit 3
It is converted into a search formula in an executable format. This search formula is passed to the search formula verification unit 4 and is judged whether it is valid or not.
Details of the processing in the search expression verification unit 4 will be described later. If it is a valid search expression, it is passed to the search expression evaluation unit 3 and a document having an element satisfying the specified condition is searched. Reference numeral 5 is a document type management unit, 6 is a data dictionary, and 7 is a database.

The search expression verification unit 4 has a verification control unit 8, a correspondence table holding unit 9, a correspondence table generation unit 10, and a reachability judgment unit 11.

The verification control unit 8 is an element that governs the whole,
The correspondence table holding unit 9, the correspondence table generation unit 10, and the reachability determination unit 11 are called as appropriate.

The correspondence table holding unit 9 has three elements: an element (starting element), an element that can appear as a child of the starting element (adjacent element set), and a set of elements reachable from the starting element (reachable element set). A correspondence table having a set as an entry is held. Here, being able to reach a certain element B from a certain element A means that the element B is a subordinate (descendant) of the instance of the element A.
Say that can appear. FIG. 3 shows a correspondence table generated from the document type shown in FIG. For example, a "section" is adjacent to the element "article" as a subordinate element, and any of the elements "section", "heading", and "paragraph" can be reached from the element "article". It is shown that.

The correspondence table generator 10 generates the above correspondence table with reference to the structure of the document type for the designated document type.

The reachability determination unit 11 scans the correspondence table (see FIG. 3) held in the correspondence table holding unit 9 when the two elements of the source and the destination are given,
Check whether the destination has a neighbor element set or reachable element set of the entry whose source element is the starting element. A destination may appear as a child of the source when the destination is included in the neighboring element set. When a destination is included in the reachable element set, the destination can appear as a descendant of the source.

4 to 9 show the flow of the processing executed by the search expression verification unit 4 to verify the given search expression. This embodiment will be described along this flow.

FIG. 4 is an overall flow of verification of a search expression. The input of this processing is the search formula generated by the search formula generation unit 2. The verification control unit 8 calls the document type management unit 5 and acquires information on the schema (document type) that is the search target of the input search formula (step 6-1). Then, the correspondence table generation unit 10 is called to create a correspondence table of the schema (step 6-2).

FIG. 5 is a flow chart of the correspondence table creation process (see step 6-2 in FIG. 4). This processing is performed by the correspondence table generation unit 10. The input is schema information.
The schema information is represented by a directed graph as shown in FIG. First, the root of the input schema is selected (step 7-2). Next, a set of element types reachable from the root is obtained (step 7-3). The route is added to the variable S to the set of element types reachable from the route, and the set is held (step 7-4). Note that step 7-4
The return value in indicates the result obtained by the immediately preceding process. Then, it is checked whether the variable S has an unprocessed node (step 7-5). If there are no unprocessed nodes, the process ends (step 7-10). If there are unprocessed nodes,
Select one node (step 7-6). Obtain a set of element types adjacent to the selected node (step 7-
7). Further, a set of element types reachable from the selected node is obtained (step 7-8). Set the selected node (element type), the adjacent element type set obtained in step 7-7, and the reachable element type set obtained in step 7-8 to 3
It is passed to the correspondence table holding unit 9 as a set and the entry is registered in the correspondence table (step 7-9). After this, step 7-5
Return to

FIG. 7 is a flow chart of a process for obtaining a reachable element type set (see step 7-3 in FIG. 5). This processing is also performed by the correspondence table generation unit 10. The input of this process is an element type, and the output is a set of element types reachable from the input element type. In this flow, a variable S holding an element type set and a variable Q holding an element type queue are used. The initial value of the variable S is an empty set (step 8-
2). The initial value of the variable Q is a queue consisting of all nodes adjacent to the input node (step 8-3). First, it is determined whether the length of the variable Q is 0 (step 8-
4). If the length of the variable Q is 0, the set of elements reachable from the input element type is stored in the variable S, and the control is returned using this as a return value (step 8-1).
0). If the length of the variable Q is 1 or more, the head element of the variable Q is taken out (step 8-11). If the extracted element is included in the variable S, the process returns to step 8-4. If the fetched element is not included in S, it is checked whether it is an element type (step 8-7). If it is an element type,
The element type is added to S (step 8-8). All the nodes adjacent to the fetched element are added to the end of the variable Q (step 8-9), and the process returns to step 8-4.

FIG. 8 is a flow chart of the processing for obtaining the adjacent element type set (see step 7-7 in FIG. 5). This processing is also performed by the correspondence table generation unit 10. The input of this process is an element type, and the output is a set of element types adjacent to the input element type. Similar to the process of obtaining the reachable element type set, this flow also uses the variable S holding the element type set and the variable Q holding the element type queue. The initial value of the variable S is an empty set (step 9-2). The initial value of the variable Q is
It is a queue consisting of all nodes adjacent to the input node (step 9-3). First, it is determined whether the length of the variable Q is 0 (step 9-4). If the length of the variable Q is 0, the set of elements reachable from the input element type is stored in the variable S, and the control is returned using this as a return value (step 9-8). If the length of the variable Q is 1 or more, the head element of the variable Q is taken out (step 9-5).
It is checked whether the fetched element is an element type (step 9)
-6). If it is an element type, the element type is added to the variable S (step 9-7), and the process returns to step 9-4. If it is not the element type, all the nodes adjacent to the fetched element are added to the end of the variable Q (step 9-9), and step 9-
Return to 4.

FIG. 9 shows the verification process of the retrieval expression node (FIG. 4).
(See Step 6-4). This process
The reachability determination unit 11 performs this. The input of this process is the node of the search expression, and the output is the truth value indicating whether the node is valid or not. First, the correspondence table held in the correspondence table holding unit 9 is scanned to find an entry having the input node as the starting element type (step 10-2). Next, it is checked whether there is an unprocessed adjacent node (step 1
0-3). If all processing is completed, the return value will be true,
The control is returned (step 10-12). If there is an unprocessed adjacent node, select one node (step 10-
4). It is determined whether the selected node is designated as a child of the input node (step 10-
5). If specified as a child, the selected node is
It is checked whether the entry is included in the neighboring element type set (step 10-7). If it is not included, the return value is set to false and the control is returned (step 10-7). If it is included, the selected node is verified (step 10-
8). If the selected node is not designated as a child in step 10-5, that is, if it is designated as a descendant, it is checked whether or not the selected node is included in the reachable element type set of the entry (step 10- 1
1). If it is not included, the return value is set to false and the control is returned (step 10-13). If so, go to step 10-8. If the verification result in step 10-8 is false, the return value is set to false and control is returned (step 10-
10). If true, return to step 10-3.

In the present embodiment, the correspondence table is constructed every time the retrieval formula is verified. However, the correspondence table is constructed at the time of registering the document type in the database, and the table is scanned during the verification. May be.

[0035]

As described above, according to the present invention, the document type is compared with the condition specified by the search expression using the parent-child relationship and the grandchild relationship, and it is determined whether the condition is valid.

As a result, it becomes easy to determine whether the search result was not obtained due to a semantic error in the search expression or there was no instance that met the condition. In addition, the system does not have to evaluate a search expression that cannot have a search result, and waste of calculation time can be prevented.

[Brief description of drawings]

FIG. 1 is a configuration of an embodiment of a document database management device of the present invention.

FIG. 2 is an example of a graphical user interface for a query editor.

FIG. 3 is an example of a correspondence table. This is the document type correspondence table shown in FIG.

FIG. 4 is a flow of verification of a search expression.

FIG. 5 is a flowchart of a correspondence table creation process.

6 is a representation of the document type of FIG. 11 by a directed graph.

FIG. 7 is a flowchart of a process for obtaining a set of element types reachable from a certain element type.

FIG. 8 is a flowchart of a process for obtaining a set of element types adjacent to a certain element type.

FIG. 9 is a flow of a verification process of a search expression node.

FIG. 10 is an example of a document instance.

FIG. 11 is an example of a document type. This is the document type of the document instance in FIG.

FIG. 12 is an example of designation of a search target.

FIG. 13 is a configuration of a conventional document database management device.

FIG. 14 is an example of an invalid search expression. The document type used in this search formula is that shown in FIG.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Inquiry editor, 2 ... Search expression generation unit, 3 ... Search expression evaluation unit, 4 ... Search expression verification unit, 5 ... Document type management unit, 6 ...
Data dictionary, 7 ... Database, 8 ... Verification control unit, 9 ...
Correspondence table storage unit, 10 ... Correspondence table generation unit, 11 ... Reachability determination unit

Claims (2)

[Claims] 1. A document database in which a structured document that defines a structure that a document can take by a syntax rule is a management target, and a search target is specified by using a condition regarding an element of the document and a condition regarding a connection relation between the elements. A document database management device, characterized in that the management device is provided with means for matching a syntax rule with a given search expression. 2. A correspondence table in which a starting element, elements adjacent to the starting element in the lower order, and elements that can exist in the lower order of the starting element are combined into one set is generated based on a structure that the document can take. 2. The document database management apparatus according to claim 1, further comprising means and means for scanning the correspondence table based on a search expression to verify whether or not the search expression is valid.