JP2002014985A - Document search system and search document registration control method - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To improve the search efficiency of a frequently updated document file by automatically updating the contents of the document file in a database efficiently. When a document file to be registered is read (step S31), it is determined whether or not the document file has already been registered in the database (step S32). If the document file to be registered is a registered document file, an identifier indicating the registration position of the corresponding old index data in the database relating to the registered document file is obtained (step S34). Then, new index data is generated from the document file to be registered, and the index data is registered in the database together with link information to the document file to be registered (step S35). After this,
The old index data and the link information associated therewith are deleted from the database 116 (step S36).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a document retrieval system such as a natural language retrieval system and a retrieval document registration control method used in the retrieval system.

[0002]

2. Description of the Related Art Conventionally, a keyword search system is well known as a document search system. This keyword search system searches for a document including a keyword input by a user. In order to search for a target document in this system, the user needs to specify an appropriate keyword contained in the target document. In this case, especially when there is a large number of documents in the database, a large number of candidate documents can be obtained only by specifying a single keyword. Therefore, the user narrows down the candidates by additionally inputting a keyword. You need to perform a refined search.

[0003] As one using such a search method,
2. Description of the Related Art A full-text search system has been developed which searches for a document file required by a user from a large number of document files. In this document search system, in order to narrow down search candidates, it is necessary for a user to connect several words with "or", "and", or "not", create a search formula, and perform a search accordingly. .

[0004] Therefore, recently, a natural language search system capable of efficiently searching registered document data in everyday language (natural language) has been developed. In a natural language search system, simply inputting a query sentence (query) in everyday language (natural language) automatically analyzes the content of the query, syntax analysis, morphological analysis, keyword weighting processing, similarity calculation And the like, whereby an appropriate document can be easily retrieved from the database.

In such a natural language search system, a high-level document analysis process is required. Therefore, a document file to be searched is analyzed in advance, search data corresponding to the document file is generated, and the search data is stored in a database. It is necessary to register in.

[0006]

Conventionally, however, such a document registration operation is manually performed by a user, and a document file to be newly included as a search target is designated, and the registration process is started. It was necessary to perform the operation of doing. Although there are some that have an automatic registration function that automatically registers document files in a predetermined directory, there are many that consider only additional registration of document files, and already have Normally, sufficient consideration is not given to the processing when a registered document file is updated. This is because the conventional natural language search system was developed on the premise that only completed document files are to be searched.

Therefore, if a new document file exists in a certain directory, the search data corresponding to the new document file can be automatically added to the database, but the contents of the registered document file are updated. In such a case, it was necessary to manually perform the work for reflecting the updated contents in the database. Further, even when automatic re-registration is performed, search data for the same document file is double-registered again in the database, thereby increasing the data size of the search data in the database and retrieving similar search data. There has been a problem that the redundancy of data is increased, and as a result, the search efficiency is reduced.

In the future, documents that are frequently updated, such as documents in the work course of each individual in the office, as represented by groupware such as e-mail, electronic document circulation, workflow, and schedule management. It is anticipated that a mechanism for effectively utilizing files will be required.

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and enables an automatic update of the contents of a document file to be automatically reflected in a database, and efficiently updates the contents of a frequently updated document file. An object of the present invention is to provide a searchable document search system and a search document registration control method.

[0010]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a document search system for searching for a document file registered in a database. Automatically registering means for periodically executing at the time of registering a document file by the automatic registering means, means for detecting whether a document file to be newly registered is already registered in the database, If it is detected that the target document file is a registered document file, search data corresponding to the newly registered target document file is generated, and the generated search data is stored in a corresponding document in the database. Overwriting registration means for overwriting and registering the file search data. That.

According to this document search system, the registration of a document file in the database is automatically performed periodically according to a schedule specified by the user. At the time of this automatic registration, it is checked whether or not the document file to be newly registered has already been registered in the database. If it is detected that the document file to be newly registered is a document file that has already been registered, search data corresponding to the newly registered document file to be newly registered is generated, and the generated search file is generated. Data is registered over the search data of the corresponding document file in the database. By using such functions of automatic registration and overwrite registration, it is possible to automatically reflect the updated contents of the document file in the database without increasing the data size of the search data, thereby improving search efficiency. Can be achieved.

If it is detected that the document file to be newly registered is a document file that has already been registered, it is determined whether the document file to be newly registered has been updated with respect to the registered document file. May be further provided based on the file update date and time, and when the document file to be newly registered has not been updated, execution of the overwrite registration process for the document file to be newly registered by the overwrite registration means may be prohibited. preferable. As a result, the overwrite registration process for only the document file that is actually updated can be realized, and the efficiency of the registration process can be improved.

When it is detected that the document file to be newly registered is an unregistered document file, a document file similar to the document file to be newly registered is searched for using the search data in the database. By
Means for determining whether a document file having a predetermined similarity or higher has already been registered is further provided. If a document file having a predetermined similarity or higher is already registered, a search file corresponding to the document file is registered. Preferably, only the link information for the document file to be newly registered is registered in the database in association with the data. This makes it possible to share search data between similar document files, so that duplication of search data is eliminated and search efficiency can be improved.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 shows a configuration of a client server type computer system to which a document search system according to an embodiment of the present invention is applied. This computer system comprises a server computer 11 and a plurality of client computers 12, as shown. The server computer 11 and the client computer 12 are connected via a computer network such as a LAN and the Internet.

The server computer 11 includes a natural language search system 111, a management function unit 112,
Registration function unit 113, file storage device 114, and H
A TTP server 115 and the like are provided. The natural language search system 111 is a search system that can efficiently search registered document data using everyday words (natural languages). Query), and then perform parsing, morphological analysis, keyword weighting, similarity calculation, and other processing.
Search for an appropriate document from the registered documents. The natural language search system 111 includes a database 116, a natural language search engine 117, and a natural language search server 118.
It is composed of

The database 116 stores search data (index data) corresponding to a document file registered as a search target. The natural language search engine 117 is a software module for performing the arithmetic processing for the natural language search described above, and searches for a registered document that matches the inquiry by using search data in the database 116. Natural language search server 1
Reference numeral 18 denotes a software module for providing a search service to the client computer 12 via the HTTP server 115.

The management function unit 112 is a software module for operating and managing the natural language search system 111, and performs processes such as schedule management for automatic document registration and setting of an automatic registration mode. Registration function unit 113
Is a software module for periodically registering a document file in the database 116 according to a schedule specified by the user. The software module targets a document file stored in the file storage device 114 under the control of the management function unit 112. Execute the automatic registration process. The schedule specified by the user means, for example, a date and time specification relating to registration such as “automatic registration at 9:00 every morning” and “automatic registration every Monday”, and the setting is performed by a system administrator or the like.

The file storage device 114 stores various document files to be searched. The user of each client computer 12 accesses the server computer 11 through the browser 121,
Various document files such as mail, schedule, and circulation can be freely written and updated in the file storage device 114.

That is, as shown in FIG. 2, the file storage device 114
2, the user of each client computer 12 can freely perform a file input / output operation on the file storage device 114 through the browser 121. Therefore, not only a new document file is added to the file storage device 114, but also an update operation for an existing document file in the file storage device 114 is frequently performed. Since the document file in the file storage device 114 is automatically and periodically registered in the database 116 of the natural language search system 111, each user of the client computer 12 can use the natural language search server 118 through the browser 121.
By accessing the target file, the target document can be easily searched from the document files in the file storage device 114.

Here, a flow of a series of processes in the present system will be described with reference to a flowchart of FIG.

First, a request for obtaining the database name of the database 116 is issued from the management function unit 112 to the natural language search server 118, and a database name to be registered is specified based on the obtained database name. (Step S11). Subsequently, the location of the document file to be registered in the natural language search system 111 is selected by the management function unit 112 (step S12). In the present embodiment, a directory (path) specifying the file storage device 114 is selected as the location of the document file. Thereafter, based on the settings made by the system administrator, the management function unit 112 sets the automatic registration date and time for the registration function unit 113 and the registration mode in that case (step S13). Various options are available in the registration mode, such as "overwrite registration of all document files",
“Overwrite registration only for a specific document file by comparing file update date and time”, “additional registration only for link information based on similarity”, and the like can be designated. Specific contents of these registration modes will be described later with reference to FIG.

When the registration start date and time specified by the management function unit 112 is reached, the registration function unit 113
Makes a connection to the database 116 specified by
Subsequently, according to the registration mode designated by the management function unit 112, a process for registering the document file in the file storage device 114 in the database 116 is performed (step S14).

The user of each client computer 12 sends a natural language search server 118 through a browser 121.
To display the homepage of the natural language search server 118 on the display monitor of the client computer 12. Then, by inputting a query in a natural language on an input form provided on the homepage, a search process based on the query is executed by the natural language search system 111, and the search result is displayed on the screen by the browser 121. (Step S1
5). The search result includes several lines of text information in the corresponding document file and link information to the document file. When the link information is clicked, the corresponding document file in the file storage device 114 is stored in the file. It can be read from the device 114 and displayed.

It is also possible to perform a process of inputting all the text of a certain document file as a kind of inquiry sentence and searching for a document file similar to the text.

Next, referring to FIG.
An example of the registration mode selection screen provided by the above will be described. This registration mode selection screen is for allowing the user to select the contents of the registration processing in the automatic registration processing. As shown in the figure, a check box of "do not check (overwrite registration of all files)" A check box, a check box for "check / not check similarity", an input field for "similarity", and the like are displayed.

When the check box of “not checked (overwrite registration of all files)” is checked, the “overwrite registration of all document files” mode is selected. In this case, “date check” and “similarity check” are not performed. In this “overwrite registration of all document files” mode, search data corresponding to all document files in the file storage device 114 is generated, and the search data is stored in the database 11.
6 is a mode for registering in the database 116
The existing search data is overwritten by the newly created search data for the document file registered in. This is shown in FIG.

In FIG. 5, at the time of first registration (registration 1)
, The file name A,
Six document files B, C, D, E, and F are stored, and these document files A, B, C, D, E, and F are stored.
Is registered in the database 116, and it is assumed that the file G has been newly added to the file storage device 114 at the time of this registration (registration 2). . Some of the document files A, B, C, D, E, and F have been updated since the previous registration, and some have not been updated.

The search data is replaced from INDEX DATA1 to INDEX DATA2. INDE
X DATA2 includes document files A, B, C, D, E,
Existing search data composed of search data newly generated from F and G, respectively, and corresponding to document files A, B, C, D, E and F is deleted. As described above, in the “overwrite registration of all document files” mode, all the indexes of the database 116 are replaced according to the latest contents of the file storage device 114. Since the existing search data is deleted, the search data will not be duplicated no matter how many times the automatic registration process is executed.

When the "check date" check box is checked, information on the file update date and time of the registered document file and the current document file is compared, and the current document file is compared with the registered data. The above-described overwriting registration process is executed only when the data is new. The specific procedure of this processing will be described later with reference to FIG.

When a check box of "check similarity" is checked, a registered document file similar to a new document file (file G in FIG. 5) which has not been registered is checked, and a similarity input field is checked. If a similar document file having a similarity degree equal to or more than the specified degree has already been registered, the search data of the similar document file is diverted as search data for a new document file. The specific procedure of this process will be described with reference to FIG.

As described above, by setting the registration mode selection screen shown in FIG. 4, the following four types of registration modes can be selectively used.

Registration mode # 1: Overwrite registration of all document files (no check) Registration mode # 2: Date check Registration mode # 3: Similarity check Registration mode # 4: Date and similarity check Next, a flowchart of FIG. Refer to the registration function section 1
The procedure of the automatic registration process executed by the server 13 will be described. When the registration date and time specified by the management function unit 112 comes, the registration function unit 113 first connects to the registration destination database 116 through the natural language search server 118 (step S21). Next, the registration function unit 113
Reads the first document file from the file storage unit 114 (step S22), determines the type of the current registration mode (step S23), and performs the process of registering the document file in the database 116 according to the determination result. (Steps S24 to S27). Thereafter, the registration function unit 113 determines whether or not the registration processing has been completed for all the document files in the file storage unit 114 (step S28). If there is an unprocessed document file, the processing from step S22 is performed. Is executed again. Since the type of the registration mode is common to all the document files, the chucking process in the registration mode may be omitted when registering the second and subsequent document files.

(Registration Mode # 1: Overwrite Registration of All Document Files) Next, a specific procedure of the registration process in the registration mode # 1 will be described with reference to the flowchart of FIG.

The registration function unit 113 stores the file storage unit 11
When a document file to be registered is read from Step 4 (Step S31), first, it is determined whether or not the document file has already been registered in the database 116 (Step S32). This is performed by checking whether or not search data corresponding to a document file having the same file name as the document file to be registered has already been registered in the database 116. If the document file to be registered is not registered, the registration function unit 113 generates search data (index data) by extracting a text from the document file and analyzing the text, and generates the search data. It is registered in the database 116 together with the link information to the document file (step S3)
3). Actually, the search data includes not only the index but also part or all of the extracted text data, but what information is registered as the search data depends on the design of the search system. Therefore, it is simply referred to as search data or index data below.

If the document file to be registered is a registered document file, the registration function unit 113 executes an overwriting process. That is, the registration function unit 113 acquires an identifier (pointer information of a record, etc.) indicating the registration position of the corresponding old index data on the database 116 regarding the document file already registered (step S34), A new index data is generated by extracting a text from the document file and analyzing the text, and the index data is registered in the database 116 together with link information to the registration target document file (step S35).
Then, thereafter, the registration function unit 113 deletes the old index data and the link information associated therewith from the database 116 using the acquired identifier (step S36). This is shown in FIG. FIG. 8 shows an example in which the file A to be registered has already been registered in the database 116. The index data relating to the file A to be registered and the ring information to the file A are newly registered in the database 116, and the registered index data relating to the file A and the ring information to the file A are deleted from the database 116.

The processing shown in FIG. 7 is individually performed for each document file to be registered. Thus, the content of the database 116 can be updated to the latest one without wasting index data.

(Registration mode # 2: date check)
With reference to the flowchart of FIG. 9, a specific procedure of the registration process in the registration mode # 2 will be described.

The registration function unit 113 stores the file storage unit 11
When the document file to be registered is read from Step 4 (Step S41), first, it is determined whether or not the document file is already registered in the database 116 (Step S42). If the document file to be registered is not registered,
The registration function unit 113 generates search data (index data) by extracting text from the document file and analyzing the text, and stores the search data together with link information to the document file in the database 1.
16 (step S43).

On the other hand, when the document file to be registered is a registered document file, the registration function unit 113
By checking the file update date and time, it is determined whether or not the overwriting process is necessary (step S44). That is, the file update date corresponding to the registered document file on the database 116 is compared with the file update date of the document file to be registered, and the document file to be registered is newer than that on the database 116. Is determined. Of course, the file update date corresponding to the registered document file is
16 may be managed for each document file, but the date and time when the previous automatic registration process was executed is used as the file update date corresponding to the registered document file, and is compared with the file update date of the document file to be registered. You may do it.

If the file update date of the document file to be registered is newer than the file update date corresponding to the registered document file in the database 116, the same procedure as in steps 34 to S36 in FIG.
Search data and link information corresponding to the content of the document file to be registered are overwritten and registered (step S
45) If the file update date of the document file to be registered is equal to or older than the file update date corresponding to the registered document file on the database 116, the process for overwrite registration in step S45 is skipped. This is shown in FIG.

In FIG. 10, at the time of the first registration (registration 1), four document files having file names A, B, C and D are stored in the file storage device 114, and these document files A and B are stored. , C, and D are registered in the database 116. At the time of this registration (registration 2), the file E is newly added to the file storage device 114, and It is assumed that the files B and C have been updated after the first registration (registration 1). Search data is INDEX DATA1
From INDEXDATA2. INDE
X DATA2 is composed of search data and link information respectively corresponding to the document files A, B, C, D, and E. The search data and link information newly generated this time are files B, C, and E, respectively. . For files A and D, existing search data and link information are diverted as they are. The existing search data and link information of the files B and C are deleted.

(Registration mode # 3: similarity check) Next, referring to the flowchart of FIG.
3 will be described.

The registration function unit 113 stores the file
When a document file to be registered is read from Step 4 (Step S51), first, it is determined whether or not the document file is already registered in the database 116 (Step S52). If the document file to be registered is not registered,
The registration function unit 113 acquires the user-specified similarity specified in the similarity input field (step S53).
Next, the registration function unit 113 extracts a whole text from the document file to be registered and inputs it to the natural language search system 111 as a query sentence, thereby executing a search for a document similar to the document to be registered (step). S
54).

If there is no search result, the registration function unit 113 generates search data (index data) by extracting a text from the document file to be registered and analyzing the text. Data is stored in the database 11 together with link information to the document file.
6 (step S55).

If there is a search result, the registration function unit 113 transmits the natural language search system 111 together with the search result.
The value of the similarity output from is checked, and the magnitude is compared with the similarity specified by the user (step S56). If the similarity of the search document is lower than the user-specified similarity,
The registration function unit 113 performs the process of step S55,
If the similarity of the search document is equal to or greater than the user-specified similarity, only the link information to the registration target document file is registered in the database 116, and the search data corresponding to the registration target document file is generated and registered. Is not performed (step S57). This is shown in FIG.

FIG. 12 shows an example in which the document file to be newly registered is the document file G, and the document file A having a similarity higher than the similarity designated by the user to the document file G is the document file A. . In this case, the generation and registration of the search data for the document file G are not performed, and the link information to the document file G is additionally registered while being associated with the search data (index data) of the document file A. In this case, as a search result for a certain query (query), for example, as shown in FIG. 13, a search result including a part of a similar document corresponding to the query and link information indicating document files A and G is displayed. Will be done.

If it is determined in step S52 in FIG. 11 that the document file to be registered is a registered document file, the registration function unit 113 executes the same procedure as in steps S34 to S36 in FIG. Then, overwrite registration of search data and link information corresponding to the content of the document file to be registered is performed (step S58).

(Registration Mode # 4: Date and Similarity Check) Next, a specific procedure of the registration process in the registration mode # 4 will be described with reference to the flowchart of FIG. This example is an example in which the similarity check processing described in FIG. 11 and the date check processing described in FIG. 9 are used in combination.

That is, when the registration function unit 113 reads a document file to be registered from the file storage unit 114 (step S61), it first determines whether or not the document file has already been registered in the database 116 (step S61). S62). If the document file to be registered has not been registered, the registration function unit 113 acquires the user-specified similarity specified in the similarity input field (step S6).
3). Next, the registration function unit 113 extracts a whole text from the document file to be registered and inputs it to the natural language search system 111 as a query sentence, thereby executing a search for a document similar to the document to be registered (step). S64).

If there is no search result, the registration function unit 113 generates search data (index data) by extracting a text from the document file to be registered and analyzing the text. Data is stored in the database 11 together with link information to the document file.
6 (step S65).

If there is a search result, the registration function unit 113 transmits the natural language search system 111 together with the search result.
The value of the similarity output from is checked, and the magnitude is compared with the similarity specified by the user (step S66). If the similarity of the search document is lower than the user-specified similarity,
The registration function unit 113 performs the process of step S65,
If the similarity of the search document is equal to or greater than the user-specified similarity, only the link information to the registration target document file is registered in the database 116, and the search data corresponding to the registration target document file is generated and registered. Is not performed (step S67).

If it is determined in step S62 that the document file to be registered is a registered document file, the registration function unit 113 checks the file update date and time to determine whether the overwriting process is necessary. The presence or absence is determined (step S68). That is, the file update date corresponding to the registered document file on the database 116 is compared with the file update date of the document file to be registered, and the document file to be registered is newer than that on the database 116. Is determined. Of course, as described above, the date and time when the automatic registration process was executed last time may be used as the file update date corresponding to the registered document file, and may be compared with the file update date of the document file to be registered.

If the file update date of the document file to be registered is later than the file update date corresponding to the registered document file on the database 116, the same procedure as in steps 34 to S36 in FIG.
Search data and link information corresponding to the content of the document file to be registered are overwritten and registered (step S
69) If the file update date of the document file to be registered is the same as or older than the file update date corresponding to the registered document file on the database 116, the process for overwrite registration in step S69 is skipped.

As described above, in the present embodiment, the automatic document registration function is applied to the natural language search system 111, and all document files are overwritten and registered (registration mode # 1) as the registration mode at the time of automatic document registration. ), Overwrite registration based on presence / absence of file update by date check (registration mode # 2), registration of link information only based on presence / absence of similar document by similarity check (registration mode # 3),
And registration (registration mode # 4) by a combination of date check and similarity check is prepared, so that even if registration processing by automatic registration is repeatedly executed, the same search data is registered twice or triple. The updated contents of the document file can be automatically and efficiently reflected in the database 116 without causing the problem that the document file is updated. In particular, since the process of registering the document file included in the file storage device 114 commonly used by the plurality of client computers 12 is automated, the unnecessary data size of the database 116 does not increase. It is also possible to effectively use a document file that is frequently updated, such as a document in a work course of each individual in the office.

In the above description, the case where the document search system according to the present embodiment is applied to a client server type computer system has been described as an example. However, the present invention is similarly applied to a stand-alone computer. be able to.

Further, since the natural language retrieval system 111 of this embodiment and the functions for its automatic registration control (management function unit 112 and registration function unit 113) are realized by computer programs, the computer programs are read by a computer. By recording the program on a possible storage medium, the same effect as in the present embodiment can be obtained only by introducing a computer program into a normal computer through the storage medium.
The computer program can be distributed not only through a storage medium but also through a communication medium.

The present invention is not limited to the above-described embodiment, and can be variously modified at the stage of implementation without departing from the scope of the invention. Further, the embodiments include inventions at various stages, and various inventions can be extracted by appropriately combining a plurality of disclosed constituent elements. For example, even if some components are deleted from all the components shown in the embodiment, at least one of the problems described in the column of the problem to be solved by the invention is to be solved.
Can be solved and the effect (at least one of them) described in the section of the effect of the invention can be obtained, a configuration from which this component is deleted can be extracted as the invention.

[0058]

As described above, according to the present invention,
The updated content of the document file can be automatically and efficiently reflected in the database, and the document file that is frequently updated can be efficiently searched.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a client server type computer system to which a document search system according to an embodiment of the present invention is applied.

FIG. 2 is an exemplary view for explaining an example of an operation mode of the document search system according to the embodiment.

FIG. 3 is an exemplary flowchart illustrating a flow of a series of processes in the document search system of the embodiment.

FIG. 4 is an exemplary view showing an example of a registration mode selection screen provided by a management function unit provided in the document search system of the embodiment.

FIG. 5 is an exemplary view for explaining the principle of overwrite registration processing applied to the document search system of the embodiment.

FIG. 6 is an exemplary flowchart illustrating the procedure of an automatic registration process executed by the document search system of the embodiment.

FIG. 7 is an exemplary flowchart showing the procedure of a specific registration process in a registration mode # 1 executed by the document search system of the embodiment.

FIG. 8 is a view for explaining the principle of the overwriting process in FIG. 7;

FIG. 9 is an exemplary flowchart showing the procedure of a specific registration process in a registration mode # 2 executed by the document search system of the embodiment.

FIG. 10 is a view for explaining the principle of overwrite registration processing based on the presence or absence of file update in FIG. 9;

FIG. 11 is an exemplary flowchart showing the procedure of a specific registration process in a registration mode # 3 executed by the document search system of the embodiment.

FIG. 12 is a view for explaining the principle of link information registration processing in FIG. 11;

FIG. 13 is a diagram showing an example of a search result when the link information registration processing in FIG. 11 is used.

FIG. 14 is an exemplary flowchart showing the procedure of a specific registration process in a registration mode # 4 executed by the document search system of the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... Server computer 12 ... Client computer 111 ... Natural language search system 112 ... Management function part 113 ... Registration function part 114 ... File storage device 115 ... HTTP server 116 ... Database 117 ... Natural language search engine 118 ... Natural language search server

Claims (12)

[Claims] 1. A document retrieval system for retrieving a document file registered in a database, comprising: an automatic registration unit for periodically registering a document file in the database on a schedule specified by a user; Means for detecting whether a document file to be newly registered is already registered in the database, and detecting that the document file to be newly registered is a registered document file Is generated, search data corresponding to the document file to be newly registered is generated,
A document search system comprising overwriting registration means for overwriting and registering the generated search data with respect to the search data of the corresponding document file in the database. 2. When it is detected that the document file to be newly registered is a document file that has already been registered, whether the document file to be newly registered has been updated with respect to the registered document file Means for determining whether or not the document file to be newly registered has not been updated, wherein the overwriting registration processing for the document file to be newly registered by the overwriting registration means is performed when the document file to be newly registered has not been updated. 2. The document search system according to claim 1, wherein execution is prohibited. 3. When it is detected that the document file to be newly registered is an unregistered document file, a document file similar to the document file to be newly registered is searched for by using search data in the database. Means for determining whether or not a document file having a predetermined similarity or higher is already registered by retrieving; if a document file having a predetermined similarity or higher is already registered, 2. The document search system according to claim 1, wherein only link information for the document file to be newly registered is registered in the database in association with search data corresponding to the file. 4. A document retrieval system for retrieving a document file registered in a database, comprising: an automatic registration unit for periodically registering a document file in the database according to a schedule specified by a user; Means for detecting whether or not a document file to be newly registered is already registered in the database; and detecting that the document file to be newly registered is an unregistered document file. In this case, it is determined whether or not a document file having a predetermined similarity or higher is already registered by searching for a document file similar to the document file to be newly registered using the search data in the database. Means, and if a document file having a predetermined similarity or higher is not registered, Generate search data corresponding to the document file to be recorded, register the generated search data and link information to the newly registered document file in the database, and register a document file having a predetermined similarity or higher. And a means for registering, in the database, only link information for the document file to be newly registered in association with search data corresponding to the document file. 5. A search document registration control method applied to a document search system for searching for a document file registered in a database, wherein the registration of the document file in the database is periodically executed according to a schedule specified by a user. Automatically registering, and when registering a document file in the automatic registration step, detecting whether or not a document file to be newly registered is already registered in the database; and If it is detected that the document file has already been registered, generate search data corresponding to the document file to be newly registered,
An overwriting registration step of overwriting and registering the generated search data over the search data of the corresponding document file in the database. 6. When it is detected that the document file to be newly registered is a document file that has already been registered, whether the document file to be newly registered has been updated with respect to the registered document file Judging whether or not the document file to be newly registered has not been updated. If the document file to be newly registered has not been updated, the overwriting registration process for the document file to be newly registered by the overwriting registration step is performed. 6. The search document registration control method according to claim 5, wherein execution is prohibited. 7. When it is detected that the document file to be newly registered is an unregistered document file, a document file similar to the document file to be newly registered is searched for by using search data in the database. The method further comprises a step of determining whether or not a document file having a predetermined similarity or higher is already registered by searching, and if a document file having a predetermined similarity or higher is already registered, 6. The search document registration control method according to claim 5, wherein only link information for the document file to be newly registered is registered in the database in association with search data corresponding to the file. 8. A search document registration control method applied to a document search system for searching for a document file registered in a database, wherein the registration of the document file in the database is periodically executed on a schedule specified by a user. Automatically registering, and when registering a document file in the automatic registration step, detecting whether or not a document file to be newly registered is already registered in the database; and When it is detected that the document file is an unregistered document file, a document file similar to the document file to be newly registered is searched for by using search data in the database, and a document file having a predetermined similarity or higher is obtained. Determining whether or not is already registered; and When the document file is not registered, search data corresponding to the document file to be newly registered is generated, and the generated search data and link information for the document file to be newly registered are registered in the database. Registering only link information to the newly registered document file in association with the search data corresponding to the document file if a document file having a predetermined similarity or higher is already registered. And a method for controlling registration of a search document. 9. A computer-readable storage medium storing a computer program for searching for a document file registered in a database, wherein the computer program specifies registration of the document file in the database by a user. An automatic registration step that is periodically executed according to a schedule of; and, when registering a document file by the automatic registration step, a step of detecting whether a document file to be newly registered is already registered in the database, If it is detected that the document file to be newly registered is a document file that has already been registered, search data corresponding to the document file to be newly registered is generated,
An overwriting registration step of overwriting and registering the generated search data with respect to the search data of the corresponding document file in the database. 10. The computer program, when it is detected that the document file to be newly registered is a document file that has already been registered, the computer program deletes the document file to be registered newly from the corresponding document file. Determining whether or not the document file to be newly registered has not been updated if the document file to be newly registered has not been updated. 10. The storage medium according to claim 9, wherein execution of an overwrite registration process is prohibited. 11. The computer program, when it is detected that the document file to be newly registered is an unregistered document file, searches the database for a document file similar to the document file to be newly registered. Further comprising a step of determining whether or not a document file having a predetermined similarity or higher is already registered by searching using the application data, wherein a document file having a predetermined similarity or higher is already registered 10. The storage medium according to claim 9, wherein only the link information for the document file to be newly registered is registered in the database in association with the search data corresponding to the document file. 12. A computer readable storage medium storing a computer program for searching for a document file registered in a database, wherein the computer program specifies registration of the document file in the database by a user. An automatic registration step that is periodically executed according to a schedule of; and, when registering a document file by the automatic registration step, a step of detecting whether a document file to be newly registered is already registered in the database, When it is detected that the document file to be newly registered is an unregistered document file, a document file similar to the document file to be newly registered is searched for by using search data in the database, and thereby a predetermined file is searched. A document file with a similarity of Determining whether a document file having a predetermined similarity or more is not registered, and generating search data corresponding to the document file to be newly registered; Data and link information for the document file to be newly registered are registered in the database, and when a document file having a predetermined similarity or higher is already registered, the document file is linked to search data corresponding to the document file. Registering only link information for the document file to be newly registered in the database.