JP2003122794A - Full-text search device, full-text search method, program, and recording medium - Google Patents

Abstract

(57) [Summary] [Problem] To provide a full-text search device that shortens the response time of registration / deletion processing as viewed from a user. SOLUTION: In registration processing means 3, text division means 6 is provided.
Then, the document data is divided into partial character strings, and registration processing is performed on the document data storage unit 7 and the small-scale full-text index storage unit 9 for registration. The deletion processing unit 4 reads out the document data from the document data storage unit 7 based on the input document identifier,
The dividing unit 6 divides the document data into partial character strings, and deletes the index if it is an index of the small-sized full-text index storage unit 9 for registration; otherwise, deletes the small-sized full-text index storage unit 10 for deletion. The deletion process is performed by recording the index in the. The search processing means 5 divides the search character string into partial character strings by the division means 6 based on the input search conditions, and converts the search results in the storage unit 10 from the search results in the large-scale full-text index storage unit 8 for search and the storage unit 9. Output as a subtraction search result.
The merging means 11 performs data transfer between the storage units 8, 9, and 10.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a full-text search device, a full-text search method, a program, and a recording medium, and more particularly to a full-text search device for searching a document containing a designated character string from a plurality of document data. , Full-text search method, program, and recording medium. INDUSTRIAL APPLICABILITY The present invention can be applied to a system for managing a large amount of document data, such as a document management system, an electronic library system, and a patent publication search system.

[0002]

2. Description of the Related Art In recent years, a large amount of electronic documents and information about the documents have been distributed via the Internet or the like due to the development of information communication technology. A document retrieval apparatus has been proposed which retrieves a desired document with high accuracy and at high speed when distributing this electronic document and information.

In such a document search device, a keyword search method or a full-text search method is used. A full-text search device that uses the full-text search technique is a device that collates an arbitrary search character string with all documents to be searched and extracts documents that include the search character string without omission. As described above, a great deal of human power is not required to pre-assign keywords to all documents to be searched. Various types of full-text search devices have been proposed, and one of them is a device that adopts a transposed (index) file system. In the transposed file method, as an auxiliary file for searching, a document in which characters / words / n-grams (n-character concatenations) appear, or a transposed file that records the appearance position in those documents is constructed in advance, and the entire text is written. At the time of retrieval, the retrieval is performed using only the transposed file, which enables very high-speed retrieval, which is effective for a system that requires high-speed retrieval of a large number of documents.

For details of the full-text search method in general and the transposed file method, refer to the document "Information Search Algorithm".
(Kenji Kita, Kazuhiko Tsuda, Masamitsu Shirahori, Kyoritsu Shuppan Co., Ltd., pp.160-179), the prior art of Japanese Patent Laid-Open No. Hei 11-073429, and the full-text search system conference, 1998 activity report (/www.ftsanet.com/dbtokyo99/Db99.htm)
, Etc., and the description thereof is omitted because it is known.

As a conventional technique adopting the transposed file method, Japanese Patent No. 3024544 discloses a real-time processing data stored separately from a search index file so that the search process can be performed even while the search index file is being updated. An information retrieval device capable of performing is described. Further, Japanese Patent Application Laid-Open No. 7-146880 describes a document search device and method capable of shortening the registration time by registering a new document in a sub index smaller than the main index. .

However, in the transposed file system including the above-mentioned publications, it is necessary to construct a transposed file which is usually several times as large as the original data, and the full-text index in the transposed file system has a large amount of registered document data. Accordingly, the registration / deletion process takes time, and the response time of the registration / deletion process seen from the user side becomes long as the full-text search device.

[0007]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and is a full-text search device capable of shortening the response time of registration and deletion processing seen from the user side. A full-text search method, a program for causing a computer to function as the device, a program for causing a computer to execute the steps of the method,
It is an object of the present invention to provide a computer-readable recording medium recording these programs.

Furthermore, the present invention provides a full-text search device, a full-text search method, and a computer capable of shortening the response time of the registration and deletion processing seen from the user side and eliminating the unprocessable period. It is an object of the present invention to provide a program for causing the computer to execute the method, a program for causing a computer to execute the procedure of the method, and a computer-readable recording medium recording the program.

[0009]

According to a first aspect of the present invention, in a full-text search device for searching a document including a designated character string from a plurality of document data, a document data storage unit for storing registered document data is provided. , Full-text index storage for searching,
It has registration processing means for performing registration processing regarding document data, deletion processing means for performing deletion processing regarding document data, and search processing means for performing search processing, and a full-text index storage unit for registration is provided for the full-text for search. In addition to the index storage unit, it is characterized by further comprising a merging unit for merging data from the registration full-text index storage unit to the search full-text index storage unit.

According to a second aspect of the present invention, in the first aspect of the invention, the merging means stores the full-text index when merging data from the registration full-text index storage section to the search full-text index storage section. The feature is that processing is performed for each transposed list that is a component.

According to a third aspect of the present invention, in a full-text search device for searching a document including a designated character string from a plurality of document data, a document data storage section for storing registered document data, and a full-text index for search. A storage unit, a registration processing unit that performs registration processing related to document data, a deletion processing unit that performs deletion processing related to document data, and a search processing unit that performs search processing, and a full-text index storage unit for registration and deletion. A full-text index storage unit for searching separately from the full-text index storage unit for searching, and further, from the full-text index storing unit for registration and the full-text index storing unit for deletion, the full-text index storage unit for searching. It is characterized in that it has a merging means for merging data into a copy.

According to a fourth aspect of the present invention, in the third aspect of the invention, the merging means stores data from the full-text index storage unit for registration and the full-text index storage unit for deletion to the full-text index storage unit for retrieval. When merging is performed, the processing is performed for each transposed list that is a constituent element of the full-text index.

According to a fifth aspect of the present invention, in any one of the second to fourth aspects of the present invention, the merge means sets the number of document data registered in the full-text index storage unit for registration to a predetermined number. When it reaches, a process of merging data into the full-text index storage unit for retrieval is performed.

According to a sixth aspect of the present invention, in the invention according to any one of the second to fourth aspects, the merging means, when the capacity of the full-text index storage unit for registration reaches a predetermined capacity, The present invention is characterized by performing a process of merging data into the full-text index storage unit for retrieval.

According to a seventh aspect of the present invention, in the fifth or sixth aspect of the present invention, a plurality of registration full-text index storage units are provided, and a process of merging data into the search full-text index storage unit is performed. Another feature of the present invention is that the registration process is performed using another registration full-text index storage unit different from the registration full-text index storage unit.

According to an eighth aspect of the present invention, in the fifth or sixth aspect of the present invention, the registration full-text index storage unit is provided, and one of the registration full-text index storage units is used for the search full-text index storage unit. While the process of merging the data into the department,
Another feature is that the registration process is performed using another full-text index storage unit for registration.

According to a ninth aspect of the present invention, in the fifth or sixth aspect of the present invention, there is provided storage unit management means for managing another full-text index storage unit for registration different from the full-text index storage unit for registration. The storage unit management means newly creates another full-text index storage unit for registration when there is no full-text index storage unit for registration capable of storing a full-text index for registration during the registration process. It is what

According to a tenth aspect of the present invention, in the ninth aspect, the storage unit managing means deletes the surplus registration full-text index storage unit.

According to an eleventh aspect of the present invention, in the invention of the ninth or tenth aspect, the registration processing means stores the full-text index storage unit for searching from one full-text index storage unit for registration among the full-text index storage unit for registration. While the data merging process is being performed, another registration full-text index storage unit is used to perform the registration process.

According to a twelfth aspect of the present invention, in a full-text search device for searching a document including a designated character string from a plurality of document data, a document data storage section for storing the registered document data, and a full-text index for search. The storage unit, a registration processing unit that performs registration processing regarding document data, a deletion processing unit that performs deletion processing regarding document data, and a search processing unit that performs search processing. A full-text index storage unit for search is provided separately, and a merging unit for merging data from the full-text index storage unit for deletion to the full-text index storage unit for search is provided. .

According to a thirteenth aspect of the present invention, in the twelfth aspect of the present invention, the merging means merges the data from the deletion full-text index storage unit into the search full-text index storage unit. The feature is that processing is performed for each transposed list that is a component.

According to a fourteenth aspect of the present invention, in the invention of any one of the third to thirteenth aspects, the merge means sets the number of pieces of document data registered in the deletion full-text index storage unit to a predetermined number. When it reaches, a process of merging data into the full-text index storage unit for retrieval is performed.

According to a fifteenth aspect of the present invention, in the invention according to any one of the third to thirteenth aspects, the merging means, when the capacity of the deletion full-text index storage unit reaches a predetermined capacity, The present invention is characterized by performing a process of merging data into the full-text index storage unit for retrieval.

The invention of claim 16 is the invention of claim 14 or 15.
In the invention described above, another full text for deletion different from the full text index storage for deletion, which has a plurality of full text index storage for deletion and is performing the process of merging data to the full text index storage for search The feature is that the deletion process is performed using the index storage unit.

The invention of claim 17 is the invention of claim 14 or 15.
In the invention described above, two full-text index storage units for deletion are provided, and while performing the process of merging data from one full-text index storage unit for deletion to the full-text index storage unit for retrieval, The deletion processing is performed by using one deletion full-text index storage unit.

The invention of claim 18 is the invention of claim 14 or 15.
In the invention, there is provided a storage unit managing means for managing another full-text index storage unit for deletion different from the full-text index storage unit for deletion, and the storage unit management unit has a full-text index for deletion in the deletion processing. When there is no deletion full-text index storage unit capable of storing the index, another deletion full-text index storage unit is newly created.

According to a nineteenth aspect of the present invention, in the eighteenth aspect of the present invention, the storage unit managing means deletes an excess full-text index storage unit for deletion.

The invention of claim 20 is the invention of claim 18 or 19.
In the invention described above, while the deletion processing unit is performing a process of merging data from one deletion full-text index storage unit of the deletion full-text index storage unit to the search full-text index storage unit, The deletion processing is performed by using the deletion full-text index storage unit.

According to a twenty-first aspect of the present invention, a document data storage unit for storing registered document data, a search full-text index storage unit for storing a search full-text index, and a registration process for performing a registration process on the document data. Full-text search method in a full-text search system for searching a document including a specified character string from a plurality of document data A full-text index storage unit for registration, which is provided separately from the full-text index storage unit for search, and an index storage step of storing a full-text index for registration, and the search from the full-text index storage unit for registration. And a merging step of merging the data into a full-text index storage unit for use in the.

According to a twenty-second aspect of the present invention, in the merging step of the twenty-first aspect, when the data is merged from the registration full-text index storage unit to the search full-text index storage unit, The feature is that processing is performed for each transposed list that is a component.

According to a twenty-third aspect of the present invention, a document data storage unit for storing the registered document data, a search full-text index storage unit for storing a search full-text index, and a registration process for performing a registration process for the document data. Full-text search method in a full-text search system for searching a document including a specified character string from a plurality of document data In the above, the full-text index storage section for registration and the full-text index storage section for deletion, which are provided separately from the full-text index storage section for search, respectively store a full-text index for registration and a full-text index for deletion. And a merge step of merging data from the registration full-text index storage unit and the deletion full-text index storage unit to the search full-text index storage unit. Is obtained by it said.

According to a twenty-fourth aspect of the present invention based on the twenty-third aspect, the merging step stores data from the full-text index storage unit for registration and the full-text index storage unit for deletion to the full-text index storage unit for retrieval. When merging, the processing is performed for each transposed list that is a constituent element of the full-text index.

The invention of claim 25 is based on claims 22 to 24.
In any one invention of the above, the merging step comprises:
When the number of document data items registered in the full-text index storage unit for registration reaches a predetermined number, a process of merging data into the full-text index storage unit for search is performed. is there.

The invention of claim 26 is based on claims 22 to 25.
In any one invention of the above, the merging step comprises:
When the capacity of the full-text index storage unit for registration reaches a predetermined capacity, a process of merging data into the full-text index storage unit for search is performed.

The invention of claim 27 is based on claim 25 or 26.
In the invention, the full-text search system has a plurality of full-text index storage units for registration, and the full-text search method is a full-text index for registration performing a process of merging data into the full-text index storage unit for search. It is characterized in that the registration process is performed by using another registration full-text index storage unit different from the storage unit.

The invention of claim 28 is the invention of claim 25 or 26.
In the invention, the full-text search system has two full-text index storage units for registration, the full-text search method, one of the full-text index storage unit for registration data to the full-text index storage unit for search. While performing the merging process, another registration full-text index storage unit is used to perform the registration process.

The invention of claim 29 is the invention of claim 25 or 26.
In the invention, there is a storage unit management step of managing another full-text index storage unit for registration different from the full-text index storage unit for registration, and the storage unit management step includes the full-text index storage unit for the registration processing. It is characterized by including a step of newly creating another full-text index storage unit for registration when there is no full-text index storage unit for registration capable of storing the index.

The invention of claim 30 is characterized in that, in the invention of claim 29, the storage unit managing step includes a step of deleting an excessive registration full-text index storage unit.

The invention of claim 31 is the invention of claim 29 or 30.
In the invention described above, the registration processing means may perform another process while merging data from one registration full-text index storage unit of the registration full-text index storage unit to the search full-text index storage unit. The feature is that the registration process is performed using the full-text index storage unit for registration.

According to a thirty-second aspect of the present invention, a document data storage unit for storing the registered document data, a search full-text index storage unit for storing a search full-text index, and a registration process for performing a registration process for the document data. Full-text search method in a full-text search system for searching a document including a specified character string from a plurality of document data In a separate full-text index storage unit for deletion, an index storage step for storing a full-text index for deletion in the deletion full-text index storage unit, and the search from the full-text index storage unit for deletion And a merging step of merging the data into a full-text index storage unit for use in the.

According to a thirty-third aspect of the invention, in the thirty-second aspect of the invention, the merging step is performed when the data is merged from the deletion full-text index storage unit to the search full-text index storage unit. The feature is that processing is performed for each transposed list that is a component.

The invention of claim 34 is based on claims 23 to 33.
In any one invention of the above, the merging step comprises:
When the number of pieces of document data registered in the deletion full-text index storage unit reaches a predetermined number, a process of merging data into the search full-text index storage unit is performed. is there.

The invention of claim 35 is based on claims 23 to 33.
In any one invention of the above, the merging step comprises:
When the capacity of the deletion full-text index storage unit reaches a predetermined capacity, a process of merging data into the search full-text index storage unit is performed.

The invention of claim 36 is based on claim 34 or 35.
In the invention described above, the full-text search system has a plurality of full-text index storage units for deletion, and the full-text search method is a full-text index for deletion performing a process of merging data into the full-text index storage unit for search. The deletion processing is performed by using another deletion full-text index storage unit different from the storage unit.

The invention of claim 37 is based on claim 34 or 35.
In the invention described above, the full-text search system has two full-text index storage units for deletion, and the full-text search method includes data from one deletion full-text index storage unit to the full-text index storage unit for search. While performing the merging process, another deleting full-text index storage unit is used to perform the deleting process.

The invention of claim 38 is based on claim 34 or 35.
In the invention, there is provided a storage unit management step of managing another full-text index storage unit for deletion different from the full-text index storage unit for deletion, and the storage unit management step includes the full-text index storage unit for deletion in the deletion processing. The present invention is characterized by including a step of newly creating another full-text index storage unit for deletion when there is no full-text index storage unit for deletion capable of storing an index.

According to a thirty-ninth aspect of the invention, in the thirty-eighth aspect of the invention, the storage section managing step includes a step of deleting an excessive deletion full-text index storage section.

The invention of claim 40 relates to claim 38 or 39.
In the invention described above, while the deletion processing unit is performing a process of merging data from one deletion full-text index storage unit of the deletion full-text index storage unit to the search full-text index storage unit, The deletion processing is performed by using the deletion full-text index storage unit.

The invention of claim 41 is for causing a computer to function as the full-text search device according to any one of claims 1 to 20, or for the computer to be the full-text search method according to any one of claims 21 to 40. Is a program for executing.

The invention of claim 42 is a computer-readable recording medium in which the program according to claim 41 is recorded.

[0051]

BEST MODE FOR CARRYING OUT THE INVENTION According to the present invention, a small-scale full-text index is separately prepared for registration and deletion to prevent deterioration of response time for registration and deletion, and a large-scale full-text index is searched during search processing. It is a full-text search device that adds search results of a small-scale full-text index for registration to results and excludes search results of a small-scale full-text index for deletion, and returns the search results to the user.
The method described in Japanese Patent Application No. 2001-78026 by the present applicant is applied to a full-text search device to solve the above-mentioned problems.

The above-mentioned Japanese Patent Application No. 2001-78026 discloses a database management system capable of further improving the update performance during system operation while maintaining the performance capable of responding to a high-level search request at high speed. A program and a recording medium are described, and it is characterized by increasing the throughput of registration / deletion by providing a data holding means for registration / deletion separately from the data holding means for search. However, in the method described in the above specification, the identifier of the document data registered in the small-scale index by the data transfer means from the small-scale full-text index for registration and deletion to the large-scale full-text index for search. The original document data is acquired from and registered and deleted in a large-scale index. As described above, since it takes time to register / delete the large-scale full-text index, the data transfer process takes a long time. Generally, since the search process cannot be performed during the registration / deletion process to the full-text index, there is a problem that the response time of the search process from the user's perspective becomes poor.

In the present invention, in the data transfer means from the small-scale full-text index to the large-scale full-text index, the full-text index of the transposed file system is used instead of the original document data, that is, the construction of the full-text index. By using the transposed list as an element, the time required for data transfer is shortened.

FIG. 1 is a block diagram for explaining the function of the full-text search device according to one embodiment of the present invention, and FIG. 2 is a hardware configuration example in the case where the full-text search device in FIG. FIG. 3 and FIG. 3 are diagrams showing a hardware configuration example when the full-text search device in FIG. 1 is configured by a server / client.

The full-text search device according to the present invention is a device for searching a document including a specified character string from a plurality of document data (a plurality of digitized documents). In the present specification, "full-text search" in the full-text search device means a search device for all character strings to be searched, and therefore, for example, a document with a tag such as SGML. If so, only the character string between the predetermined tags may be targeted as appropriate.

Referring to FIG. 1, in the present embodiment, the input means 1 inputs text data for registration processing, a document identifier for deletion processing, search conditions for search processing, and the like, and the registration processing means respectively. 3, the deletion processing means 4, and the search processing means 5. The registration processing means 3 performs registration processing regarding document data. The registration process in the registration processing means 3 is performed on the document data storage unit 7 and the registration small-scale full-text index storage unit 9. The deletion processing means 4 performs a deletion process regarding the document data. In the deletion processing in the deletion processing unit 4, the document data stored in the document data storage unit 7 is read based on the document identifier input by the input unit 1, and the text dividing unit 6 is used to store a small-scale full-text index storage for registration. If it is the index registered in the section 9, it is deleted, and if it is not the registered index, the index is recorded in the deletion small-scale full-text index storage section 10. The text dividing unit 6 divides the document data into partial character strings in the registration process and the partial characters from the document data in the deletion process, which are required in each of the registration processing unit 3, the deletion processing unit 4, and the search processing unit 5. Division processing into columns and division processing from a search condition (search character string) in the search processing into partial character strings are performed.

Further, the search processing in the search processing means 5 is executed by the large-scale full-text index storage section 8 for search, the small-scale full-text index storage section 9 for registration, and the small-scale full-text index storage section 10 for deletion, and stored. A result obtained by subtracting the search result in the storage unit 10 from the search results of the units 8 and 9 is obtained, and is output by the output unit 2 as the search result. In the merging means 11,
A merge process (broadly referred to as data transfer) is performed among the search large-scale full-text index storage unit 8, the registration small-scale full-text index storage unit 9, and the deletion small-scale full-text index storage unit 10.

Although not particularly described below, regarding the deletion processing in the deletion processing means 4, for example, only the document data to be deleted is deleted and the processing time is reduced without using the deletion small-scale full-text index storage unit 10. On other days such as holidays to be obtained, another method of managing the document data (and index) for deletion, such as updating the data in the search large-scale full-text index storage unit 8 to match the document data existing in the document data storage unit 7, It is also possible to employ a mode in which only the registration process using the registration small-scale full-text index storage unit 9 is performed. On the contrary, it is possible to adopt a mode in which the registration process using the registration small-scale full-text index storage unit 9 is not performed and only the deletion process using only the deletion small-scale full-text index storage unit 10 is performed.

In the stand-alone hardware configuration shown in FIG. 2, the input means 1 in FIG. 1 is realized by the input device 21, and the output means 2 is realized by the display device 22. The various processing means 3 to 6 and 11 are main control devices (CPU,
Memory, etc.) 24, and various storage units 7 to 10, for example, all as storage devices 25, or as individual storage devices,
Further, it is realized as a file in the storage device 25. For example, when performing a full-text search according to the present invention using one limited storage device, it is preferable to allocate the area to be used depending on whether the search process is mainly performed or the registration / deletion process is mainly performed. . Further, the input / output control device 23 receives the input device 2 according to the control signal from the main control device 24.
1 and the display device 22 are controlled.

In the hardware configuration of the server / client shown in FIG. 3, the input means 1 in FIG. 1 is realized by the input device 31 of the client 30, and the output means 2 is used.
Is realized on the display device 32 of the client 30. The various processing means 3 to 6 and 11 are the client 30 and the server 5.
0 of the main control units (CPU, memory, etc.) 34, 52, and various storage units 7 to 10, for example, all server 5.
No. 0 storage device 53, individual storage devices connected to the server 50, and further as files in the storage device 25. Also, the client 3
0, the network control devices 35 and 51 of the server 50
Client 30 and server 5 via network 40
Controls data transmission between 0s. Further, the input / output control device 33 of the client 30 controls the input device 21 and the display device 22 according to the control signal of the main control device 34.

Hereinafter, an example of the operation of the full-text search device according to the present embodiment configured as described above will be described in detail.
4 to 6 are flowcharts for explaining an example of processing in the full-text search device in FIG. When the full-text search device receives a processing request from the user (step S1), first, is the processing a registration processing (step S)?
2) Whether it is deletion processing (step S3) or search processing (NO in step S3) is determined. The full-text search device will execute the following processes based on this determination.

(Registration Processing) In order to execute the registration processing, the user first creates document data and registers the document data from the input means 1. The registration processing means 3 saves the document data in the document data storage unit 7, and at the same time, determines an identifier (document identifier) indicating the document data (step S1).
1). For example, in the case of a document with a tag such as SGML, only a character string between predetermined tags may be appropriately targeted. Further, in the registration processing means 3, the text dividing means 6 is used to obtain the partial character string (token) and the appearance position information of the token from the document data (step S12).
Finally, the document identifier and the appearance position information of each token are recorded in the registration small-scale full-text index storage unit 9 (step S1).
3). The “record” here is a record in the full-text index of the storage unit (the same applies hereinafter), and the processing such as step 13 is also called an index storage step. Note that the dividing method used by the text dividing means 6 may be a method of using N character sets as tokens, or a method of performing morphological analysis and using words as tokens. In the following example, only the text segmentation means using the method of using N character sets as tokens will be described, but the method of applying the words obtained by morphological analysis to tokens can be similarly applied. Further, as will be described later, after the recording in step S13, timely merge processing is performed (steps S14 to S16).

FIG. 7 is a diagram for explaining the processing in the full-text search device of FIG. 1, showing an example of the full-text index. The full-text index of the transposed file system will be described in detail with reference to the example of FIG. Registered document data as Document 1 and Document 2
The contents (here, the contents obtained by dividing the text by the text dividing means 6) are denoted by reference numeral 6 in FIG.
1, 62. Here, the numbers on the left side of each document represent the number of characters from the beginning of the character string. In other words, in document 1, "full text search" is the 11th character from the beginning,
"Method" is the 20th and 60th characters, "Full-text search method" is 31
It means that it appears in the letter. In Document 2, the “search method” is the first character from the beginning, and the “method” is 24
The first character, "full text", means that it appears at the 30th and 42nd characters.

When the two-character set is a partial character string,
All the partial character strings in the document are extracted, and the appearance positions (the number of characters from the beginning) in those documents are collected for each partial character string and recorded in the index. For example, from document 1 "full text"
Since "11" and "31" appear at positions "Sentence" appearing at positions 12 and 32, they are recorded in the index. In the index, not only the appearance position in the document but also the document identifier for identifying in which document it appears and the number of appearances are added and recorded.
The format is as shown by reference numeral 63 in FIG. For example,
Transposed list {1,2, (11,3
1)} and {2,2, (30,42)} appear twice in document 1 and their positions are 11 and 31, respectively, and appear twice in document 2 and their position is 3
It means 0,42.

(Delete Process) To execute the delete process, the user first inputs the document identifier of the document to be deleted from the input means 1. Next, the deletion processing means 4 reads the document data corresponding to the document identifier from the document data storage unit 7 (step S21). Further, in the deletion processing means 4,
The partial character string (token) and the appearance position information of the token are obtained from the document data by using the text dividing means 6 (step S22). For example, in the case of a document with a tag such as SGML, only a character string between predetermined tags may be appropriately targeted. It is determined whether the document identifier is a document identifier registered in the registration small-scale full-text index (step S23). If the document identifier is registered in the registration small-scale full-text index, appearance position information of each token is determined. Is deleted from the registration small-scale full-text index storage unit 9 (step S25). If the document identifier is not registered in the small-scale full-text index for registration (if registered in the large-scale full-text index for search),
The document identifier and the appearance position information of each token are recorded in the deletion small-scale full-text index storage unit 10 (step S24). Then, the deletion processing means 4 deletes the document data corresponding to the document identifier from the document data storage unit 7 (step S29). Also, as will be described later, after the recording in step S24, a merge process is performed at a proper time (step S26).
~ S28).

(Search Process) To execute the search process, the user first inputs a search character string from the input means 1. Next, in the search processing means 5, the text division means 6 is used to obtain a token from the search character string (step S4).
Further, the search processing means 5 uses the search large-scale full-text index of the search large-scale full-text index storage unit 8 to obtain a set (Rs) of document identifiers of document data including a search character string (step S5). Small-scale full-text index storage unit 9 for registration
Using the registration small-scale full-text index for registration, a set (Ri) of document identifiers of document data including a search character string is obtained (step S6). Furthermore, the retrieval processing means 5 uses the deletion small-scale full-text index of the deletion small-scale full-text index storage unit 10,
A set of document identifiers of document data (R
d) is obtained (step S7). The search processing means 5 performs the following set operation on the obtained set of document identifiers (Rs, Ri, Rd) and sets the result as the search result (R) (step S8), and outputs it to the user through the output means 2. A set of document identifiers of document data including the search character string is output (step S9). R = Rs + Ri−Rd However, + is a logical sum operator and − is a logical difference operator.

The search process will be described in detail by taking the full-text index 63 of FIG. 7 as an example. When the search character string is "full text search", the text segmentation means extracts three tokens "full text", "text check", and "search". Next, the three transposed lists of corresponding tokens in full-text index 63 are examined. It is found that “full-text search” exists at the 11th and 31st characters of the document identifier 1 when searching for a token having a difference of 1 in the appearance position.

(Merging Process) The merging process by the merging means 11 is a process which is a substitute for the data transferring means in the above-mentioned Japanese Patent Application No. 2001-78026. Compared to the case where registration / deletion processing is performed using the original document data, the transposition list already created at the start of processing is directly used, so the time required to cut out tokens by text segmentation processing and create that transposition list is unnecessary. Therefore, the data transfer time can be shortened. In the present invention, the data transfer processing (data transfer step) is also called merge processing (merge step) because it is processing between transposed lists. By registering / deleting the document data in the full-text search device between transposed lists, the transposed list that has already been created is directly used when registering / deleting data in the full-text index for search. It is possible to shorten the time for the merge processing to the full-text index and shorten the waiting time for the search processing.

In order to execute the merge processing, first, for all tokens of the registration small-scale full-text index, (a) processing for extracting the transposed list of the token from the full-text index (step S14), and (b) search. A process of adding the previous transposed list to the end of the transposed list of the corresponding token of the large-scale full-text index (step S15) is performed. Next, the small scale full-text index for registration is emptied (step S16). In addition, for all tokens of the small-scale full-text index for deletion, (c) processing for extracting an inverted list of the token from the full-text index (step S26), and (d) of corresponding tokens of the large-scale full-text index for search A process of deleting the appearance position information in the transposed list extracted in (c) from the transposed list (step S
27) is performed. Next, the small scale full-text index for deletion is emptied (step S28).

FIG. 8 is a diagram for explaining the outline of the merge process by taking the transposed list of the token “full text” of the full text index 63 in FIG. 7 as an example. Transposed list of full-text index for search 7
As 1, the transposed list {1, 2, (1
1,31)}, {2,2, (30,42)}, the transposed list 72 of the full-text index for registration, the transposed lists {5,2, (4,16)}, {8,1 for "full text" , (3)}
The case where the merge processing 73 is performed will be described. Merge process 73
By executing the transposed list {1, 2, (11, 31)}, {2, 2, (30, 4
2)}, {5, 2, (4, 16)}, {8, 1,
(3)} (74) is obtained. Further, the transposed list for the "full text" is obtained by merging the transposed list with the transposed list {1, 2, (11, 31)} for the "full text" as the transposed list 76 of the full text index for deletion. {2,2, (30,42)}, {5,2,
(4, 16)}, {8, 1, (3)} (77) are obtained.

(Form 1 of merge process) The merge process is performed when the number of document identifiers registered in the small scale full-text index for registration in the small scale full-text index storage unit 9 for registration reaches a predetermined number. Is started by the registration processing means 3 and executed by the merging means 11.

(Form 2 of merge processing) The merge processing is started by the registration processing means 3 when the storage capacity (size) in the registration small-scale full-text index storage unit 9 reaches a predetermined size. It may be executed by the merging means 11. With this mode, when used as an application mode in which the size of the document data registered by the user varies, when the small document data is continuously registered, the small-scale full-text index for registration is added. It is possible to prevent the merge process from being started before the registration time becomes long. By setting the size as the start condition, the merging processing time can be equalized. Further, in the case of the above-described merge processing (mode 1), the number of cases is set as the starting condition, and it is not necessary to manage the size of the full-text index storage unit, so that there is an advantage that the processing is simplified.

(Form 3 of Merge Processing) The merge processing of the small scale full-text index for deletion is started by the deletion processing means 4,
It may be executed by the merging means 11. The starting condition may be when the number of document identifiers registered in the small scale full-text index for deletion reaches a predetermined number.

(Form 4 of merge processing) The merge processing of the small scale full text index for deletion is started by the deletion processing means 4,
It may be executed by the merging means 11. The activation condition may be when the size of the small scale full text index storage unit for deletion 10 reaches a size designated in advance. The forms 3 and 4 have an advantage that the time of the merge process can be shortened when the deletion process does not often occur.

By each mode of the merge process as described above,
With the full-text search device, it is possible to start the merge processing of the full-text index under conditions suitable for the characteristics of the document data to be registered / deleted and the characteristics of the field of use, reducing the number of times merge processing occurs and improving the overall system throughput. It becomes possible. Further, the start condition of the merge may be variable depending on the time required for the merge process, or the merge caused by the registration and the merge caused by the deletion may be simultaneously activated under any of the start conditions. .

In the full-text search device according to the above-described embodiment, the method described in Japanese Patent Application No. 2001-78026 by the present applicant is applied to a full-text search device using a transposed file type full-text index, and the scale is small. In the data transfer method from a full-text index to a large-scale full-text index, the time required for data transfer is shortened by using the transposed list that is a component of the transposed file method full-text index instead of using the original document data. There is.

A full-text search device according to another embodiment of the present invention, which will be described below, is the full-text search device according to the above-described embodiment, which is the Japanese Patent Application No. 2001-101024 filed by the present applicant.
Write delay database management method described in the specification,
The method used in the device, the program, and the recording medium is applied. As a result, while the data transfer (merge processing of the transposed list) from the small-scale full-text index for registration or deletion to the large-scale full-text index for search is being performed, the small-scale full-text index for registration or deletion is executed. It is possible to solve the problem that the storage unit cannot be used and the registration process or the deletion process cannot be executed.

FIG. 9 is a block diagram for explaining the function of the full-text search device according to another embodiment of the present invention. In the full-text search device according to the present embodiment, two small-scale full-text indexes for registration and deletion are prepared, and while the merge (data transfer) to the large-scale full-text index is being performed, the other small-scale full-text index is deleted. By executing the registration process or the deletion process using the index, the unprocessable period is eliminated.
That is, in the full-text search device according to the present embodiment,
By providing two small scale full-text indexes for registration, it is possible to perform registration processing even while merge processing is in progress, and by providing two small scale full-text indexes for deletion, deletion processing can be performed even during merge processing. It will be possible. According to the present embodiment, for example, when a document is read by a scanner or the like and OCR processing is performed to register each document, it is suitable when registration processing and merge processing by the registration processing are frequently and continuously performed. With such image data, full-text search can be performed with high response, as with normal application data.

The registration small-scale full-text index storage unit 9 described in FIG. 1 has two storage units, a registration small-scale full-text index storage unit A (9a) and a registration small-scale full-text index storage unit B (9b). The deletion small-scale full-text index storage unit 10 described in FIG. 1 is the deletion small-scale full-text index storage unit A (10a).
The present embodiment will be described as having two storage units, namely, a small-scale full-text index storage unit for deletion B (10b). Note that the hardware configuration examples described with reference to FIGS. 2 and 3 can also be applied to the full-text search device according to this embodiment.
However, one or more of these storage units may be replaced by the storage device 25,
It is effective to provide it on the memory instead of 53.

Hereinafter, an example of the operation of the full-text search device according to the present embodiment configured as described above will be described in detail.
10 to 12 are flowcharts for explaining a processing example in the full-text search device in FIG. 9. Full-text search device
When a processing request is received from the user (step S3
1) First, whether the process is a registration process (step S32) or a deletion process (step S33),
It is determined whether it is a search process (NO in step S33). The full-text search device will execute the following processes based on this determination.

(Registration Processing) In order to execute the registration processing, the user first creates document data and registers the document data from the input means 1. The registration processing means 3 saves the document data in the document data storage unit 7, and at the same time, determines an identifier (document identifier) indicating the document data (step S4).
1). Further, in the registration processing means 3, the text dividing means 6 is used to obtain a partial character string (token) and appearance position information of the token from the document data (step S4).
2). The dividing method and full-text index used by the text dividing unit 6 are as described above. The document identifier and the appearance position information of each token are used to store the small-scale full-text index storage unit for registration at that time (for example, the small-scale full-text index storage unit A for registration.
(9a)) (step S43).

After the recording in step S43, the merging process is performed at a proper time, but here, it is assumed that the merging process is performed based on the merging start condition. First, step S4
It is determined whether the merge start condition is satisfied as a result of recording in step 3 (step S44). If the merge start condition is not satisfied, the process ends. It should be noted that each form of the condition for starting the merge process described in the embodiment of FIGS.
It is also applicable to this embodiment. It is also determined whether the other registration small-scale full-text index storage unit (here, registration small-scale full-text index storage unit B (9b)) is executing the merge process (step S45). If it is being executed, it waits for the end of the merge processing.

If the merge start condition is satisfied and the other registration small-scale full-text index storage unit B (9b) is not executing merge processing, the registration small-scale full-text index storage unit A (9b).
The merging process (steps S47 to S49) described later is started for the small scale full-text index A for registration in (9a), and the storage unit to be recorded for the next registration process is the small scale full-text index storage unit for registration. Switching from A (9a) to the other small scale full-text index storage unit for registration B (9b) (step S
46). When the merging process is started, the merging means 11
Executes the merge processing asynchronously with the registration processing means 3.

(Delete Process) To execute the delete process, the user first inputs the document identifier of the document to be deleted from the input means 1. Next, the deletion processing means 4 reads the document data corresponding to the document identifier from the document data storage unit 7 (step S51). Further, in the deletion processing means 4,
The partial character string (token) and the appearance position information of the token are obtained from the document data using the text dividing means 6 (step S52).

Next, it is judged whether the document identifier is a document identifier registered in the small scale full-text index for registration (step S5).
3) If the document identifier is registered in the registration small-scale full-text index, the appearance position information of each token is deleted from the registration small-scale full-text index storage unit 9 (9a and 9b) (step S55). . If the document identifier is not registered in the small-scale full-text index for registration (if it is registered in the large-scale full-text index for search), the document identifier and the appearance position information of each token are stored in the small-scale full-text for deletion at that time. It is recorded in the index storage unit (for example, the small-scale full-text index storage unit for deletion A (10a)) (step S54). Then, the deletion processing means 4
In step S62, the document data corresponding to the document identifier is deleted from the document data storage unit 7 (step S62).

After the recording in step S54, the merging process is performed at a proper time, but here it is assumed that the merging process is performed based on the merging start condition. First, step S5
As a result of recording in step 4, it is determined whether or not the merge start condition is satisfied (step S56). If the merge start condition is not satisfied, the process ends (the process of step S62 is required). It should be noted that each form of the conditions for starting the merge process described in the embodiments of FIGS. 1 to 8 can be applied to this embodiment. On the other hand, the deletion small-scale full-text index storage unit (here, the deletion small-scale full-text index storage unit B (10
It is also determined whether b)) is executing the merge process (step S57). If it is being executed, it waits for the end of the merge processing.

When the merge start condition is satisfied and the other delete small-scale full-text index storage unit B (10b) is not executing merge processing, the deletion in the delete small-scale full-text index storage unit A (10a). The later-described merge processing (steps S59 to S61) is started for the small scale full-text index A for use, and the storage section to be recorded for the next delete processing is deleted from the small scale full-text index storage section A (10a). The other small scale full-text index storage unit B (10b) for deletion is switched (step S58). When the merge processing is activated, the merge means 11 executes the merge processing asynchronously with the delete processing means 4.

(Search Process) To execute the search process, the user first inputs a search character string from the input means 1. Next, the search processing means 5 uses the text dividing means 6 to obtain a token from the search character string (step S3).
4). Further, the search processing means 5 uses the search large-scale full-text index of the search large-scale full-text index storage unit 8 to obtain a set (Rs) of document identifiers of document data including a search character string (step S35). The search processing means 5 uses the registration small-scale full-text index A of the registration small-scale full-text index storage unit A (9a).
Is used to obtain a set (RiA) of document identifiers of document data containing a search character string, and the small-scale full-text index storage unit B for registration is used.
Using the registration small-scale full-text index B in (9b), a set (RiB) of document identifiers of document data including a search character string is obtained (step S36). Further, the search processing means 5 obtains a set (RdA) of document identifiers of document data including a search character string by using the deleting small-scale full-text index A of the deleting small-scale full-text index storage unit A (10a). Using the deleting small-scale full-text index B of the deleting small-scale full-text index storage unit B (10a), a set (RdB) of document identifiers of the document data including the search character string is obtained (step S37).

The retrieval processing means 5 performs the following set operation on the obtained set of document identifiers (Rs, RiA, RiB, RdA, RdB) and sets the result as the retrieval result (R) (step S38), The output means 3 outputs a set of document identifiers of document data including the search character string to the user (step S39). R = Rs + RiA + RiB-RdA-RdB However, + is a logical sum operator and-is a logical difference operator.

(Merge Processing) In order to execute the merge processing of the registration small-scale full-text index, all of the registration small-scale full-text indexes (here, the registration small-scale full-text index A) that are the target of the merge processing are executed. For a token, (a) processing for extracting the inverted list of the token from the full-text index (step S
47), and (b) a process of adding the previous transposed list to the end of the transposed list of the corresponding token of the large-scale full-text index for search (step S48). Next, the small scale full-text index A for registration is emptied (step S49).

In order to execute the merge process of the delete small full-text index, all tokens of the delete small full-text index (here, delete small full-text index A) that is the target of the merge process are executed. , (C) processing for extracting the inverted list of the token from the full-text index (step S59), and (d) appearance position in the inverted list extracted in (c) from the inverted list of the corresponding token of the large-scale search full-text index. A process of deleting information (step S60) is performed. Next, the deletion small-scale full-text index A is emptied (step S61).
Note that the example of the merge processing of the transposed list described with reference to FIG. 8 can be applied to this embodiment.

In the embodiment described with reference to FIGS. 9 to 12, a full-text search device using three or more registration small-scale full-text index storage sections and / or three or more deletion full-text index storage sections. The form is illustrated as the next embodiment with reference to FIGS. 13 to 16.

FIG. 13 is a block diagram for explaining the function of the full-text search device according to another embodiment of the present invention.
In the full-text search device according to the present embodiment, three or more small-scale full-text indexes for registration and deletion are prepared (exemplified as three each), and the other two small-scale full-text indexes become large-scale full-text indexes. While the merge (data transfer) is being performed, the unprocessable period is eliminated by executing the registration process or the deletion process using another small-scale full-text index. That is, in the full-text search device according to the present embodiment, by providing a plurality of registration small-scale full-text indexes, another registration processing is performed even when the merge processing is performed on a plurality of registration small-scale full-text indexes. It is possible to perform the registration process even if it is open. Also, by providing multiple small scale full-text indexes for deletion, even if the merge processing is performed for multiple small scale full-text indexes for registration, other deletion processing can be performed. It is possible to perform the deletion process even if
Note that, in reality, the time required for registration and deletion is shorter than the merge time, and thus the merge processing often overlaps.

The small-scale full-text index storage unit 9 for registration described in FIG. 1 is a small-scale full-text index storage unit A (9a) for registration, a small-scale full-text index storage unit for registration B (9b), and a small-scale full-text index for registration. As a storage unit having three storage units C (9c), the deletion small-scale full-text index storage unit 10 described in FIG.
This embodiment has three storage units: a small scale full-text index storage unit for deletion A (10a), a small-scale full-text index storage unit for deletion B (10b), and a small-scale full-text index storage unit for deletion C (10c). The form will be described. Note that the hardware configuration examples described with reference to FIGS. 2 and 3 can also be applied to the full-text search device according to this embodiment. However, it is effective to provide one or more of these storage units on the memory instead of the storage devices 25 and 53.

Hereinafter, an example of the operation of the full-text search device according to the present embodiment configured as described above will be described in detail.
14 to 16 are flowcharts for explaining a processing example in the full-text search device in FIG. The full-text search device receives a processing request from the user (step S7).
1) First, whether the process is a registration process (step S72) or a deletion process (step S73),
It is determined whether it is a search process (NO in step S73). The full-text search device will execute the following processes based on this determination.

(Registration Processing) In order to execute the registration processing, first, the user creates document data and registers the document data from the input means 1. The registration processing means 3 stores the document data in the document data storage unit 7, and at the same time, determines an identifier (document identifier) indicating the document data (step S8).
1). Further, in the registration processing means 3, the partial character string (token) and the appearance position information of the token are obtained from the document data by using the text dividing means 6 (step S8).
2). The dividing method and full-text index used by the text dividing unit 6 are as described above. The document identifier and the appearance position information of each token are used to store the small-scale full-text index storage unit for registration at that time (for example, the small-scale full-text index storage unit A for registration.
(9a)) (step S83).

After the recording in step S83, the merging process is performed at a proper time. Here, it is assumed that the merging process is performed based on the merging start condition. First, step S8
It is determined whether the merge start condition is satisfied as a result of recording in step 3 (step S84). If the merge start condition is not satisfied, the process ends. It should be noted that each form of the condition for starting the merge process described in the embodiment of FIGS.
It is also applicable to this embodiment. Also, it is determined whether the other registration small-scale full-text index storage unit (here, registration small-scale full-text index storage unit B (9b)) is executing merge processing (step S85). If it is being executed, it is determined whether or not the third registration small-scale full-text index storage unit (here, registration small-scale full-text index storage unit C (9c)) is being executed (step S86). It is also determined at the same time whether the storage units B and C are executing the registration process. If the process is being executed in step S86, the end of the process is awaited. Note that, hereinafter, only the case where the most likely merge process is being executed will be described.

The merge start condition is satisfied, and any other small scale full-text index storage unit for registration B (9b) / C (9
When c) is not executing the merge process, steps S47 to S4 in FIG. 11 are performed on the registration small-scale full-text index A in the registration small-scale full-text index storage unit A (9a).
A merge process similar to that in step 9 (steps S89 to S91) is started, and a storage unit to be recorded for the next registration process is stored in the registration small-scale full-text index storage unit A (9a) as another registration small-scale full-text. Switch to the index storage unit B (9b) / C (9c) (the same storage unit as the storage unit that has not executed the merge process, and the same expression will be used hereinafter) (steps S87 / S8).
8). When the merge process is activated, the merge unit 11 executes the merge process asynchronously with the registration processing unit 3.

(Delete Process) To execute the delete process, the user first inputs the document identifier of the document to be deleted from the input means 1. Next, the deletion processing means 4 reads the document data corresponding to the document identifier from the document data storage unit 7 (step S101). Further, the deletion processing means 4 obtains the partial character string (token) and the appearance position information of the token from the document data by using the text dividing means 6 (step S102).

Next, it is judged whether the document identifier is a document identifier registered in the small scale full-text index for registration (step S1).
03), if the document identifier is registered in the registration small-scale full-text index, the appearance position information of each token is deleted from the registration small-scale full-text index storage unit 9 (9a, 9b and 9c) (step S105). If the document identifier is not registered in the small-scale full-text index for registration (if it is registered in the large-scale full-text index for search), the document identifier and the appearance position information of each token are stored in the small-scale full-text for deletion at that time. Index storage unit (for example, small-scale full-text index storage unit for deletion A (1
0a)) (step S104). Then, the deletion processing means 4 deletes the document data corresponding to the document identifier from the document data storage unit 7 (step S11).
4).

After the recording in step S104, the merging process is performed at a proper time. Here, it is assumed that the merging process is performed based on the merging start condition. First, step S
As a result of recording in 104, it is determined whether or not the merge start condition is satisfied (step S106). If the merge start condition is not satisfied, the process ends (the process of step S114 is necessary). It should be noted that each form of the conditions for starting the merge process described in the embodiments of FIGS. 1 to 8 can be applied to this embodiment. The other small-scale full-text index storage unit for deletion (here, small-scale full-text index storage unit for deletion B
It is also determined whether (10b)) is executing the merge process (step S107). If it is being executed, it is determined whether or not the third registration small-scale full-text index storage unit (here, registration small-scale full-text index storage unit C (10c)) is being executed (step S108). It is also determined at the same time whether the storage units B and C are executing the registration process. Step S1
If the process is being executed in 08, the end is waited for. Note that, hereinafter, only the case where the most likely merge process is being executed will be described.

A merge small-size full-text index storage unit B (10b) / C (1
0c) is not executing the merge process, the small scale full-text index for deletion A in the small scale full-text index storage for deletion A (10a) is processed from step S59 to step S59 in FIG.
Merge processing similar to S61 (steps S111 to S11)
3) is started, and the storage unit to be recorded for the next deletion process is deleted from the small scale full text index storage unit for deletion A (10a) to another small scale full text index storage unit for deletion B (10b) / C ( 10
Switching to c) (steps S109 / S110). When the merge processing is activated, the merge means 11 executes the merge processing asynchronously with the delete processing means 4.

(Search Processing) The search processing according to this embodiment is basically the same as the search processing described with reference to FIG. 10, and steps S34 to S39 in FIG.
Correspond to steps S74 to S79, respectively. However, in step S76, the search processing means 5 uses the registration small-scale full-text index C of the registration small-scale full-text index storage unit C (9c) in addition to the sets RiA and RiB to include document data including a search character string. Get a set of document identifiers (RiC) of. Further, in step S77, the search processing means 5
In addition to the sets RdA and RdB, a set of document identifiers of document data (R
dC) is obtained. The search processing means 5 obtains a set of the obtained document identifiers (Rs, RiA, RiB, RiC, RdA, Rd.
The following set operation is performed on (B, RdC) and the result is set as the search result (R) (step S78). R = Rs + RiA + RiB + RiC−RdA−RdB−
RdC However, + is a logical sum operator and − is a logical difference operator.

In the embodiment of FIGS. 9 to 12 or the embodiment of FIGS. 13 to 16, there is provided a full-text search device using a plurality of registration small-scale full-text index storage units and / or a plurality of deletion full-text index storage units. As described above, these full-text index storage units (other than the search large-scale full-text index storage unit) are assigned to individual storage areas in the storage device 25 or 35 or the memory described in FIG. 2 and FIG. Alternatively, the storage device 2
5 or 35 or a mode applicable when positioned as individual files stored in a memory will be illustrated as the next embodiment with reference to FIGS. 17 to 20.

FIG. 17 is a block diagram for explaining the function of the full-text search device according to another embodiment of the present invention.
In the full-text search device according to the present embodiment, one small-scale full-text index for registration and one for deletion are prepared in advance, and while the small-scale full-text index is performing merge (data transfer) to the large-scale full-text index. For example, if there is no registration (/ deleting) full-text index storage unit that can store the registration (/ deleting) full-text index during registration (/ deletion) processing, another small-scale full-text index By newly creating and executing the registration process or the deletion process, the unprocessable period is eliminated. That is, in the full-text search device according to the present embodiment, by providing a plurality of registration small-scale full-text indexes in a timely manner, even if the merge processing is performed on a plurality of registration small-scale full-text indexes, another registration processing is performed. It is possible to perform the registration process even when the merge process is performed, and even when the merge process is performed for multiple small scale full-text indexes for registration by providing multiple small scale full-text indexes for deletion in a timely manner. The deletion process can be performed even when another deletion process is being performed. Note that, in reality, the time required for registration and deletion is shorter than the merge time, and thus the merge processing often overlaps.

The full-text search device according to this embodiment has a storage unit managing means 12 for managing a small-scale full-text index storage unit for registration different from the small-scale full-text index storage unit for registration A (9a). To do. Further, regarding the deletion process, the storage unit management unit 12 also manages another small scale full text index storage unit for registration different from the small scale full text index storage unit A (10a) for deletion. The storage unit management unit 12 has a unit for newly creating another full-text index storage unit for registration when there is no full-text index storage unit for registration capable of storing a full-text index for registration during the registration process. Furthermore, the storage unit management means 12
Also has means for deleting a surplus (which will not be used in the next process) registration (/ deleting) full-text index storage unit.

In addition, the registration small-scale full-text index storage unit 9 described in FIG. 1 is changed from only the registration small-scale full-text index storage unit A (9a) to the registration small-scale full-text index storage units B (9b) and C (9).
c), D (9d) ,. ． ． The deletion small-scale full-text index storage unit 10 described in FIG. 1 is used only for the deletion small-scale full-text index storage unit A (10a). To delete small-scale full-text index storage unit B (10b), C (10c), D (10
d) ,. ． ． The present embodiment will be described assuming that the numbers are increased in a timely manner (in no particular order) and those are deleted in a timely manner.

By utilizing the small scale full-text index storage unit for registration that has been created / deleted at the appropriate time, the registration processing means 3 uses the small full-text index storage unit for registration to retrieve a large-scale search text from one of the full-text index storage unit for registration. While the process of merging data into the full-scale full-text index storage unit 8 (or another registration process) is being performed, the registration process is performed using another full-text index storage unit for registration. On the other hand, by using the deletion small-scale full-text index storage unit that has been created / deleted in a timely manner, the deletion processing means 4 uses one of the deletion full-text index storage units to delete the large-scale full-text index storage unit. While the process of merging data into the full-text index storage unit 8 (or another deletion process) is being performed, the deletion process is performed using another full-text index storage unit for deletion. Note that the hardware configuration examples described with reference to FIGS. 2 and 3 can also be applied to the full-text search device according to this embodiment. However, it is effective to provide one or more of these storage units on the memory instead of the storage devices 25 and 53.

Hereinafter, an example of the operation of the full-text search device according to the present embodiment configured as described above will be described in detail.
18 to 20 are flowcharts for explaining a processing example in the full-text search device in FIG. When the full-text search device receives a processing request from the user (step S1)
21) First, whether the process is a registration process (step S122) or a deletion process (step S12).
3) Is it a search process (NO in step S123)
To judge. The full-text search device will execute the following processes based on this determination.

(Registration Processing) In order to execute the registration processing, first, the user creates document data and registers the document data from the input means 1. The registration processing means 3 saves the document data in the document data storage unit 7, and at the same time, determines an identifier (document identifier) indicating the document data (step S1).
31). Further, in the registration processing means 3, a partial character string (token) is extracted from the document data by using the text dividing means 6.
And the appearance position information of the token are obtained (step S13
2). The dividing method and full-text index used by the text dividing unit 6 are as described above.

The storage unit management means 12 judges whether or not there is a small scale full-text index storage unit for registration which can be used at the present time according to an instruction from the registration processing unit 3 or at a proper time (step S133). If it does not exist, another small-scale full-text index storage unit for registration (for example, small-scale full-text index storage unit C for registration) is newly created (step S135). When there is a registration small-scale full-text index storage unit that can be used, the document identifier and the appearance position information of each token are stored in the registration small-scale full-text index storage unit (for example, registration small-scale full-text index storage unit A
(9a) / C) (steps S134 / S13)
6).

After the recording in steps S134 / S136, the merge process is performed at a proper time. Here, it is assumed that the merge process is performed based on the merge start condition. First,
As a result of recording in steps S134 / S136, it is determined whether or not the merge start condition is satisfied (step S13).
7). If the merge start condition is not satisfied, the process ends. It should be noted that each form of the conditions for starting the merge process described in the embodiments of FIGS. 1 to 8 can be applied to this embodiment. The other small-scale full-text index storage unit for registration (here, small-scale full-text index storage unit for registration B (9b)
/ A (9a)) is also executing merge processing (step S138). It is also determined at the same time whether the storage unit B / A is executing the registration process. Step S
If the process is being executed at 138, it waits for its end. Note that, hereinafter, only the case where the most likely merge process is being executed will be described.

When the merge start condition is satisfied and the other small-scale full-text index storage unit B (9b) / A (9a) for registration is not executing the merge process, the small-scale full-text index storage unit A for registration A ( 9a) / C small full-text index for registration A /
For C, the merge process (steps S140 to S142) similar to steps S47 to S49 in FIG. 11 is activated, and the storage unit to be recorded for the next registration process is the small scale full-text index storage unit A for registration. (9a) / C is switched to another small-scale full-text index storage unit for registration B (9b) / A (9a) (step S139). When the merge process is activated, the merge unit 11 executes the merge process asynchronously with the registration processing unit 3. Further, the storage unit management means 12 may delete the surplus (not planned to be used in the next process) registration full-text index storage unit at the time of the merge process or at a proper time.

(Delete Process) To execute the delete process, the user first inputs the document identifier of the document to be deleted from the input means 1. Next, the deletion processing unit 4 reads out the document data corresponding to the document identifier from the document data storage unit 7 (step S151). Further, the deletion processing means 4 uses the text division means 6 to obtain a partial character string (token) and appearance position information of the token from the document data (step S152).

Next, it is judged whether the document identifier is a document identifier registered in the small scale full-text index for registration (step S1).
53) If the document identifier is registered in the registration small-scale full-text index, the appearance position information of each token is deleted from all existing registration small-scale full-text index storage units 9 (9a, etc.). Step S155). When the document identifier is not registered in the registration small-scale full-text index (when registered in the search large-scale full-text index), it is recorded in the deletion small-scale full-text index storage unit described below.

The storage unit management means 12 determines whether or not there is a small scale full-text index storage unit for deletion which can be used at the present time according to a command from the deletion processing unit 3 or in a timely manner (step S154). If it does not exist, another small-scale full-text index storage unit for deletion (for example, small-scale full-text index storage unit C for deletion) is newly created (step S157). When there is a deleteable small-scale full-text index storage unit that can be used, the document identifier and the appearance position information of each token are used to delete the small-scale full-text index storage unit (for example, registration small-scale full-text index storage unit A
(10a) / C) (step S156 / S1)
58). Then, the deletion processing unit 4 deletes the document data corresponding to the document identifier from the document data storage unit 7 (step S175).

After the recording in steps S156 / S158, the merge process is performed at a proper time. Here, it is assumed that the merge process is performed based on the merge start condition. First,
As a result of recording in step S156 / S158, it is determined whether or not the merge start condition is satisfied (step S15).
9). If the merge start condition is not satisfied, the process ends (the process of step S175 is necessary). It should be noted that each form of the conditions for starting the merge process described in the embodiments of FIGS. 1 to 8 can be applied to this embodiment. It is also determined whether the other small-scale full-text index storage unit for deletion (here, small-scale full-text index storage unit for deletion B (10b) / A (9a)) is executing merge processing (step S1).
70). It is also determined at the same time whether the storage unit B / A is executing the registration process. If the process is being executed in step S170, the completion of the process is awaited. Note that, hereinafter, only the case where the most likely merge process is being executed will be described.

The merge start condition is satisfied, and any other small scale full-text index storage unit for deletion B (10b) / C (1
0c) is not executing the merge process, the deletion small-scale full-text index storage unit A (10a) / C for the deletion small-scale full-text index A / C is the same as steps S59 to S61 in FIG. Merge processing (step S172
~ S174) is started, and the storage unit to be recorded for the next deletion process is the small-scale full-text index storage unit for deletion A (10).
a) / C to another small-scale full-text index storage unit for deletion B (10
b) / A (10a) is switched (step S17)
1). When the merge processing is activated, the merge means 11 executes the merge processing asynchronously with the delete processing means 4. Further, the storage unit management means 12 may delete the surplus (which will not be used in the next process) deletion full-text index storage unit at the time of the merge process or at a proper time.

(Search Process) The search process according to this embodiment is basically the same as the search process described with reference to FIG. 10, and steps S34 to S39 in FIG.
Correspond to steps S124 to S129, respectively.
However, in step S126, the search processing means 5
As a set Ri, a set of document identifiers of document data including a search character string is obtained by using the small scale full-text indexes for registration in all the small scale full-text indexes for registration that currently exist.
Further, in step S127, the search processing means 5
As a set Rd, a set of document identifiers of document data including a search character string is obtained by using all the deleting small-scale full-text indexes of all the deleting small-scale full-text index storage units that are presently present.

Although the respective embodiments have been described above centering on the full-text search device of the present invention, the full-text search system may take the form of a full-text search method as described as the processing procedure in the full-text search device. Furthermore, the present invention also functions as a program for causing these full-text search devices to function, or as a program for causing them to function as respective means, or executes a program for executing these full-text search methods, or a processing procedure thereof. And a form as a computer-readable recording medium in which any one of these programs is recorded can be adopted.

An embodiment of a recording medium storing a program and data for realizing the full-text search function according to the present invention will be described. As the recording medium, specifically, a CD-R
An OM, a magneto-optical disk, a DVD-ROM, an FD, a flash memory, and other various ROMs and RAMs can be assumed, and these recording media are caused to cause a computer to execute the system functions of the embodiments of the present invention described above, and full-text search is performed. By recording and distributing a program for realizing the function of, the function can be easily realized. Then, the recording medium as described above is attached to an information processing device such as a computer and the program is read by the information processing device, or the program is stored in a storage medium included in the information processing device, and as necessary. By reading, the full-text search function according to the present invention can be executed.

[0122]

According to the present invention, since the registration processing and the deletion processing in the full-text search device are performed on the small-scale full-text index storage unit, the processing time can be kept short and the response to the user can be reduced. It is possible to shorten the time.

Further, according to the present invention, by providing a plurality of registration small-scale full-text indexes, the registration processing can be performed even while the merge processing is being executed, and by providing a plurality of deletion small-scale full-text indexes, the merge processing can be performed. It is possible to perform the deletion process even while executing.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating a function of a full-text search device according to an embodiment of the present invention.

FIG. 2 is a diagram showing a hardware configuration example in the case where the full-text search device in FIG. 1 is configured standalone.

FIG. 3 is a diagram showing a hardware configuration example when the full-text search device in FIG. 1 is configured by a server / client.

FIG. 4 is a flowchart for explaining a processing example in the full-text search device in FIG.

5 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 6 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 7 is a diagram for explaining the process in the full-text search device of FIG. 1, showing an example of a full-text index.

FIG. 8 is a diagram for explaining an outline of merge processing by taking an inverted list of tokens “full text” of the full text index in FIG. 7 as an example.

FIG. 9 is a block diagram illustrating a function of a full-text search device according to another embodiment of the present invention.

FIG. 10 is a flowchart for explaining a processing example in the full-text search device in FIG.

11 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 12 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 13 is a block diagram illustrating a function of a full-text search device according to another embodiment of the present invention.

FIG. 14 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 15 is a flowchart for explaining a processing example in the full-text search device in FIG.

16 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 17 is a block diagram illustrating a function of a full-text search device according to another embodiment of the present invention.

18 is a flowchart for explaining a processing example in the full-text search device in FIG.

19 is a flowchart for explaining a processing example in the full-text search device in FIG.

FIG. 20 is a flowchart for explaining a processing example in the full-text search device in FIG.

[Explanation of symbols]

1 ... Input means, 2 ... Output means, 3 ... Registration processing means, 4 ...
Deletion processing means, 5 ... Search processing means, 6 ... Text dividing means, 7 ... Document data storage section, 8 ... Large-scale full-text index storage section for search, 9 ... Small-scale full-text index storage section for registration, 9a ... Small registration area Scale full-text index storage unit A, 9b ... Registration small-scale full-text index storage unit B, 9c ... Registration small-scale full-text index storage unit C, 1
0 ... Delete small-scale full-text index storage unit, 10a ... Delete small-scale full-text index storage unit A, 10b ... Delete small-scale full-text index storage unit B, 10c ... Delete small-scale full-text index storage unit C, 11
... merging means, 12 ... storage section managing means, 21, 31 ... input device, 22, 32 ... display device, 23, 33 ... input / output control device, 24, 34, 52 ... main control device (CPU / memory), 25 , 53 ... storage device, 30 ... client, 3
5, 51 ... Network control device, 40 ... Network, 50 ... Server.

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Tetsuya Ikeda             1-3-3 Nakamagome, Ota-ku, Tokyo Stocks             Company Ricoh (72) Inventor Takuya Hiraoka             1-3-3 Nakamagome, Ota-ku, Tokyo Stocks             Company Ricoh (72) Inventor Hiroshi Takekawa             1-3-3 Nakamagome, Ota-ku, Tokyo Stocks             Company Ricoh (72) Inventor Kazushige Asada             1-3-3 Nakamagome, Ota-ku, Tokyo Stocks             Company Ricoh F term (reference) 5B075 ND03 ND23 ND40 NK02 NK49                       NR06 NR14 PP12 PP22 QM01                       UU06

Claims (42)

[Claims] 1. A full-text search device for searching a document containing a specified character string from a plurality of document data, a document data storage section for storing registered document data, a full-text index storage section for searching, and a document. A registration processing means for performing registration processing regarding data, a deletion processing means for performing deletion processing regarding document data, and a search processing means for performing search processing;
A merging unit that has a full-text index storage unit for registration separately from the full-text index storage unit for search, and further merges data from the full-text index storage unit for registration to the full-text index storage unit for search. A full-text search device comprising: 2. The merging means, when merging data from the registration full-text index storage unit to the search full-text index storage unit, performs processing for each transposed list that is a component of the full-text index. The full-text search device according to claim 1. 3. A full-text search device for searching a document containing a specified character string from a plurality of document data, a document data storage section for storing registered document data, a full-text index storage section for searching, and a document. A registration processing unit that performs a registration process regarding data, a deletion processing unit that performs a deletion process regarding document data, and a search processing unit that performs a search process;
A full-text index storage unit for registration and a full-text index storage unit for deletion are provided separately from the full-text index storage unit for search, and further, the full-text index storage unit for registration and the full-text index storage for deletion are provided. A full-text search device comprising a merging unit for merging data from a section to the full-text index storage section for searching. 4. The merging means is a constituent element of a full-text index when merging data from the registration full-text index storage section and the deletion full-text index storage section to the search full-text index storage section. The full-text search device according to claim 3, wherein processing is performed for each transposed list. 5. The merging means merges the data into the full-text index storage unit for search when the number of document data items registered in the full-text index storage unit for registration reaches a predetermined number. The processing is performed.
4. The full-text search device according to any one of 4 to 4. 6. The merging means, when the capacity of the full-text index storage unit for registration reaches a capacity designated in advance,
The full-text search device according to any one of claims 2 to 4, wherein a process of merging data into the full-text index storage unit for search is performed. 7. Another registration different from the registration full-text index storage unit that has a plurality of registration full-text index storage units and performs a process of merging data into the search full-text index storage unit. 7. The full-text search device according to claim 5, wherein the full-text index storage unit is used to perform registration processing. 8. The apparatus has two full-text index storage units for registration, and while performing a process of merging data from one full-text index storage unit for registration to the full-text index storage unit for search, 7. The full-text search device according to claim 5, wherein the registration processing is performed using another full-text index storage unit for registration. 9. A storage unit management unit for managing another full-text index storage unit for registration different from the full-text index storage unit for registration, the storage unit management unit for registering during the registration processing. 7. The full-text search device according to claim 5, wherein another full-text index storage unit for registration is newly created when there is no full-text index storage unit for registration capable of storing the full-text index. 10. The full-text search device according to claim 9, wherein the storage unit management unit deletes the surplus registration full-text index storage unit. 11. The registration processing means, while performing a process of merging data from one registration full-text index storage unit of the registration full-text index storage unit to the search full-text index storage unit, 11. The full-text search device according to claim 9, wherein the registration processing is performed using the full-text index storage unit for registration of. 12. A full-text search device for searching a document containing a designated character string from a plurality of document data, a document data storage section for storing registered document data, a full-text index storage section for searching, and a document. It has a registration processing means for performing a registration processing regarding data, a deletion processing means for performing a deletion processing regarding document data, and a search processing means for performing a search processing. A full-text search device, which is provided separately from the storage unit, and further has merge means for merging data from the full-text index storage unit for deletion to the full-text index storage unit for search. 13. The merging means, when merging data from the deletion full-text index storage unit to the search full-text index storage unit, performs processing for each transposed list that is a constituent element of the full-text index. The full-text search device according to claim 12. 14. The merging means merges data into the full-text index storage unit for retrieval when the number of document data registered in the full-text index storage unit for deletion reaches a predetermined number. 14. The full-text search device according to claim 3, wherein the full-text search device performs processing. 15. The merging means performs a process of merging data into the full-text index storage unit for retrieval when the capacity of the full-text index storage unit for deletion reaches a predetermined capacity. The full-text search device according to any one of claims 3 to 13, which is characterized in that. 16. Another deletion full-text index storage unit different from the deletion full-text index storage unit that has a plurality of deletion full-text index storage units and performs a process of merging data into the search full-text index storage unit. The full-text search device according to claim 14 or 15, wherein the full-text index storage unit is used to perform the deletion process. 17. The apparatus includes two deletion full-text index storage units, and while performing a process of merging data from one deletion full-text index storage unit to the search full-text index storage unit, 16. The full-text search device according to claim 14, wherein the deletion processing is performed using another full-text index storage unit for deletion. 18. A storage unit managing unit that manages another full-text index storage unit for deletion, which is different from the full-text index storage unit for deletion, and the storage unit management unit deletes the full-text index storage unit for deletion during the deletion processing. 16. A new full-text index storage unit for deletion is newly created when there is no full-text index storage unit for deletion capable of storing a full-text index.
Full text search device described. 19. The storage unit managing means deletes a surplus deletion full-text index storage unit.
Full text search device described. 20. While the deletion processing unit is performing a process of merging data from one deletion full-text index storage unit of the deletion full-text index storage unit to the search full-text index storage unit, 20. The full-text search device according to claim 18, wherein the deletion process is performed using the deletion full-text index storage unit. 21. A document data storage unit for storing registered document data, a search full-text index storage unit for storing a search full-text index, registration processing means for performing registration processing on document data, and document data. And a search processing unit that performs a search process.
In a full-text search method in a full-text search system for searching a document containing a specified character string from a plurality of document data, a full-text index storage section for registration, which is provided separately from the full-text index storage section for search, is registered. A full-text search method, comprising: an index storage step of storing a full-text index for searching; and a merging step of merging data from the registration full-text index storage section to the search full-text index storage section. 22. The merging step, when merging data from the registration full-text index storage unit to the search full-text index storage unit, performs processing for each transposed list that is a constituent element of the full-text index. The full-text search method according to claim 21, wherein: 23. A document data storage unit for storing registered document data, a search full-text index storage unit for storing a full-text index for search, a registration processing unit for performing registration processing relating to document data, and document data. And a search processing unit that performs a search process.
In a full-text search method in a full-text search system for searching a document containing a specified character string from a plurality of document data, a full-text index storage section for registration and a full-text index storage section for deletion provided separately from the full-text index storage section for searching. An index storage step for storing a full-text index for registration and a full-text index for deletion, respectively, for the full-text index storage section; And a merging step of merging data into the full-text index storage unit. 24. The merging step is a constituent element of a full-text index when merging data from the registration full-text index storage unit and the deletion full-text index storage unit to the search full-text index storage unit. 24. The full-text search method according to claim 23, wherein processing is performed for each transposed list. 25. The merging step merges data into the full-text index storage unit for search when the number of document data items registered in the full-text index storage unit for registration reaches a predetermined number. 25. The full-text search method according to claim 22, wherein processing is performed. 26. The merging step includes performing a process of merging data into the full-text index storage unit for search when the capacity of the full-text index storage unit for registration reaches a predetermined capacity. The full-text search method according to any one of claims 22 to 25. 27. The full-text search system has a plurality of full-text index storage units for registration, and the full-text search method performs a process of merging data into the full-text index storage unit for search. 27. The full-text search method according to claim 25, wherein the registration process is performed by using another full-text index storage unit for registration different from the index storage unit. 28. The full-text search system has two full-text index storage units for registration, and the full-text search method is one of the full-text index storage unit for registration to the full-text index storage unit for search. 27. The full-text search method according to claim 25, wherein another registration full-text index storage unit is used during the merge processing to perform the registration processing. 29. A storage unit management step of managing another full-text index storage unit for registration different from the full-text index storage unit for registration, the storage unit management step for registering at the time of the registration processing. 27. The full-text search according to claim 25 or 26, further comprising the step of newly creating another full-text index storage section for registration when there is no full-text index storage section for registration capable of storing a full-text index. Method. 30. The full-text search method according to claim 29, wherein the storage unit managing step includes a step of deleting an excessive registration full-text index storage unit. 31. While the registration processing means is performing a process of merging data from one registration full-text index storage unit of the registration full-text index storage unit to the search full-text index storage unit, 31. The full-text search method according to claim 29, wherein the registration process is performed using the full-text index storage unit for registration. 32. A document data storage unit for storing registered document data, a search full-text index storage unit for storing a full-text index for search, registration processing means for performing registration processing on document data, and document data. And a search processing unit that performs a search process.
In a full-text search method in a full-text search system for searching a document containing a specified character string from a plurality of document data, a full-text index storage section for deletion, which is provided separately from the full-text index storage section for search, is deleted. A full-text search method, comprising: an index storing step of storing a full-text index for use in a search; and a merging step of merging data from the delete full-text index storage into the search full-text index storage. 33. The merging step, when merging data from the deletion full-text index storage unit to the search full-text index storage unit, performs processing for each transposed list that is a constituent element of the full-text index. 33. The full-text search method according to claim 32. 34. The merging step merges data into the full-text index storage unit for retrieval when the number of document data items registered in the full-text index storage unit for deletion reaches a predetermined number. 34. The full-text search method according to claim 23, wherein processing is performed. 35. The merging step performs a process of merging data into the full-text index storage unit for search when the capacity of the full-text index storage unit for deletion reaches a predetermined capacity. 34. The full-text search method according to claim 23, wherein the full-text search method is a full-text search method. 36. The full-text search system has a plurality of full-text index storage units for deletion, and the full-text search method performs a process of merging data into the full-text index storage unit for search. 36. The full-text search method according to claim 34, wherein the deletion process is performed by using another deletion full-text index storage unit different from the index storage unit. 37. The full-text search system has two full-text index storage units for deletion, and the full-text search method is such that one of the deletion full-text index storage units transfers data to the full-text index storage unit for search. 36. The full-text search method according to claim 34, wherein another deletion full-text index storage unit is used to perform the deletion processing while the merging processing is performed. 38. A storage unit management step of managing another full-text index storage unit for deletion different from the full-text index storage unit for deletion, the storage unit management step including a storage unit for deletion during the deletion processing. 36. The full-text search according to claim 34 or 35, further comprising the step of newly creating another full-text index storage section for deletion when a full-text index storage section for deletion capable of storing a full-text index does not exist. Method. 39. The full-text search method according to claim 38, wherein the storage management step includes a step of deleting a surplus deletion full-text index storage section. 40. While the deletion processing means is performing a process of merging data from one deletion full-text index storage unit of the deletion full-text index storage unit to the search full-text index storage unit, 40. The full-text search device according to claim 38, wherein the deletion process is performed using the deletion full-text index storage unit. 41. A computer for causing a computer to function as the full-text search device according to claim 1.
Alternatively, the computer according to any one of claims 21 to 40.
A program for executing the described full-text search method. 42. A computer-readable recording medium in which the program according to claim 41 is recorded.