JP2002269136A - Document retrieval system and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable reversing display before different notation normalization by a pattern matching at an outputting side of retrieval results in retrieval that absorbs fluctuation of a different notation character string. SOLUTION: A document registration request processing means 2 performs different notation normalization of a document requested to be registered in a database 3 by a database operation request inputting means 2 to prepare an index and holds a character string before being subjected to different notation normalization in a prepared index even after the different notation normalization. In retrieving a document from the database 3, a document retrieval request processing means 5 performs different notation normalization of a retrieval character string, performs retrieval, and outputs a document hit by the retrieval and a character string that is before being subjected to different notation normalization of a character string hit in the retrieval and held in the index.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to a document search system and a program.

[0002]

2. Description of the Related Art Japanese Unexamined Patent Publication No. Hei 10-334118 discloses a technique for extracting an index by extracting a text from a document, reducing a size of an index by using a regular expression dictionary and a word index, and absorbing fluctuations of different notations. Is disclosed.

[0003] Japanese Patent Application Laid-Open No. 08-263508 discloses that a plurality of differently described character strings of a search character string are created and a full-text search is performed without an index to absorb fluctuations.

[0004]

For example, when indexing a plurality of documents in a format that can be displayed and edited only by a specific application, a full-text search is performed. Extract the full text from the target document. 2. Index the extracted full text. The procedure is as follows.

[0005] When a certain search condition is given, an index is searched using the search character string. In this method, when a search is performed using a different notation character string (for example, "interphone" or "interphone" ] Is a character string with the same meaning but a different notation).

[0006] Then, the notation normalization is performed on the extracted full-text text, and the notation character string as in the above example can be converted into a unified notation method, and an index can be added to the converted character string. . Also, by normalizing the search character string in a different notation, a search can be performed without being aware of the different notation.

However, the character string hit by this method cannot be determined as to what character string was in the original document.
Therefore, when displaying a document hit by the application and trying to reverse the display of the hit character string, it is necessary to find the exact position of the hit on the application side. It is necessary to normalize in different notation and search again, which is not efficient.

SUMMARY OF THE INVENTION It is an object of the present invention to enable reverse display before normalization of a different notation by pattern matching on the output side of a search result in a search that absorbs fluctuation of a different notation character string.

SUMMARY OF THE INVENTION It is an object of the present invention to output, as a search result, a character string that has not been normalized in a notation but not a character string that has been normalized in a different notation when a search hits.

[0010] It is an object of the present invention to eliminate the need for normalization work for a different notation when a character string hit in a search is displayed in reverse on the client side constituting the system.

[0011]

According to the first aspect of the present invention, an index of a document requested to be registered is created, the document is registered in a database, and a plurality of documents in the database are registered using the index. In a document search system that performs a search based on a search character string given as a target, the registration request document is normalized in a different notation, and the index is created. Document registration request processing means for retaining the previous character string in the created index;
Performing the search by normalizing the search string in a different notation,
A document search request processing means for outputting a document hit in the search and a character string hit in the search before the notation is normalized. is there.

Therefore, in a search that absorbs fluctuations of a character string of a different notation, inverted display before normalization of the different notation by a pattern match on the output side of the search result becomes possible.

According to a second aspect of the present invention, in the document search system according to the first aspect, the document registration request processing means extracts a text from the registration request document, The index is created by using

Therefore, it is possible to extract the text from the registration request document and create an index for the full text which has been normalized in different notation.

[0015] The invention described in claim 3 is the first or second invention.
In the document search system described in the above, the document search request processing means, for the document hit in the search, instead of the character string normalized by the different notation, the character string before the different notation normalization in the index It is characterized by outputting.

[0016] Therefore, when a search is hit, a character string before being normalized in a different notation can be output as a search result instead of a character string normalized in a different notation.

According to a fourth aspect of the present invention, in the document search system according to any one of the first to third aspects, the document search request processing means is constructed by a client server system, and is configured by the document registration request processing means and the document search request processing means. By performing the different notation normalization operation on the server, the client side reversely displays the character string hit in the search on a display without the different notation normalization.

Accordingly, the client side can eliminate the need for normalization of the notation when reversely displaying the character string hit in the search.

According to a fifth aspect of the present invention, an index of a document requested to be registered is created, the document is registered in a database, and a plurality of documents in the database are given using the index. In a program for causing a computer to execute a process of performing a search based on a search character string, the registration request document is normalized in a different notation, and the index is created. A document registration request process for holding a character string in an index after creation, and performing the search by normalizing the search character string in a different notation, and performing the search in the different notation of a document hit in the search and a character string hit in the search A program for causing a computer to execute a document search request process that outputs a character string before being normalized and a computer.

Therefore, in a search that absorbs fluctuations of a character string of a different notation, inverted display before normalization of the different notation by pattern matching on the output side of the search result becomes possible.

According to a sixth aspect of the present invention, in the program according to the fifth aspect, the document registration request processing includes extracting a text from the registration request document, and indexing the indexed text with respect to the full-text normalized. Is created.

Therefore, it is possible to extract the text from the registration request document and create an index for the full text which has been normalized in different notation.

The invention according to claim 7 is the invention according to claim 5 or 6.
Wherein the document search request process outputs a character string before the alternative notation normalization in the index in place of the alternative notation normalized character string for the document hit in the search. It is characterized by.

Therefore, when a search is hit, a character string before being normalized in a different notation but not a character string normalized in a different notation can be output as a search result.

[0025]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described.

FIG. 1 is a functional block diagram of the document search system 1 of the present invention. As shown in FIG. 1, the database operation request input means 2 comprises, for example, a client terminal for inputting an operation request such as a database 3 prepared in a server. A database operation request is input via the interface. The result of the process for the input operation request is sent again to the database operation request input means 2 and displayed on the display to notify the user. When handling a document whose contents can be displayed and edited only by a specific application, the application needs to be installed in the user's client terminal which is the database operation request input unit 2.

The database 3 registers a document requested to be registered by the database operation request input means 2, and
The index created from the document requested to be registered is stored.

An example of the processing procedure of the document registration request processing means 4 will be described with reference to the flowchart shown in FIG. As shown in FIG. 2, when a document registration request and a registered document are received (step S1), the document requested to be registered is inspected (step S2), and if full-text data can be extracted (Y in step S3). , Continue the process, otherwise end. If the full text can be extracted (Y in step S3), the document requested to be registered is stored in a database, a unique number (hereinafter, document ID) is assigned, and the full text is extracted (step S4). Then, the notation normalization is performed on the extracted full-text text to remove the fluctuation of the notation (step S5), and an index is created for the all-text normalized normalization (step S6).

"Normalization normalization" performed in step S5 is the following processing. That is, although the notation is different, the meaning is the same character string. For example, the presence or absence of long sounds like "computer" and "computer", the difference between uppercase and lowercase letters such as "ABC" and "abc", and the difference between whether or not the sound is muted like "active" and "active" , "Veranda" and "Veranda" are the differences in pronunciation notation. Then, converting this into a unique notation is referred to as a notation normalization.

The document registration request processing means 4 retains, in the index, the character string that has not been normalized after the abnormal notation normalization. For this reason, a technique is employed in which the character string before the normalization of the variant notation is recorded together with the part of the full text that has been normalized. For example, if “hex” is abnormally normalized to “hexa”, the full-text text after the abnormal notation normalization is “……”.
Hey, "K, Ku" ... At this time, "" becomes an escape character and must be treated specially in the full text. When "" is used in the full-text before the notation normalization, it is distinguished by arranging two such as "".

The index creation performed in step S6 is performed as follows. That is, an index is created for the full-text text that has been normalized in the different notation by the means. The index is analogous to a telephone directory. If a character string is a person's name, the document ID, the number of appearances, the appearance position, (if the character string has been normalized in a different notation), The character string is a telephone number. In the telephone directory, if a person's name is found, a telephone number can be found, and if a search character string is found in an index, a corresponding document ID, the number of appearances, an appearance position, (if a character string normalized in a different notation) are different. You can see the character string before notation normalization. By recording the character string before the normalization in the different notation, the character string before the normalization in the different notation can be extracted even in a full-text search using the character string after the normalization in the different notation.

An example of the processing procedure of the document search request processing means 5 will be described with reference to the flowchart shown in FIG. As shown in FIG. 3, when a search character string is received together with a document search request (step S11), normalization is performed on the search character string to absorb fluctuations (step S12). Then, full-text search is performed at high speed using the index created at the time of document registration (step S13). If there is a hit, the document ID of the document containing the hit, and if there is a character string before being normalized in a different notation, the character If no hit is found in the column, that fact is returned (step S14). Then, the application is used to perform pattern matching using the character string to cause the application to perform reverse display (step S1).
5).

FIG. 4 is a block diagram schematically showing a hardware configuration of the document search system 1. As shown in FIG. As shown in FIG. 4, the document search system 1 shows the configuration of the server, and has a CPU 12 having a function of controlling each unit.
, A ROM 13 storing a BIOS and the like, and a CPU 12
And a RAM 14 serving as a work area of the computer.

The bus 15 has a hard disk 16, an input device 17 such as a keyboard and a mouse, a CRT and an LCD.
, A storage medium reading device 20 that reads data from a storage medium 19 such as a CD, DVD, or FD, and a communication control device 22 that is connected to a network 21.

The hard disk 16 stores various programs such as a document management program for implementing the program of the present invention. This program is read from the storage medium 19 by the storage medium reading device 20, or
It is installed on the hard disk 16 by downloading from the Internet or the like via the network 21. With this installation, the document search system 1 becomes operable. Note that a program such as a document management program may be a part of a function of a specific application software. Further, it may operate on a predetermined OS.

The database 3 is built in the server by a program such as a document management program, and the functions of the document registration request processing means 4 and the document search request processing means 5 are realized in the server. The network 21 is a LAN, and is connected to the client (not shown) that implements the database operation request input unit 2.

[0037]

According to the first aspect of the present invention, it is possible to perform reverse display before normalization of different notation by pattern matching on the output side of the search result in search in which fluctuation of the different notation character string is absorbed.

According to a second aspect of the present invention, in the document search system according to the first aspect, text can be extracted from a registration request document and an index can be created for a full-text normalized in a different notation.

The third aspect of the present invention is the first or second aspect.
In the document search system described in (1), when a search is hit, a character string before being subjected to the variant notation normalization can be output as a search result, instead of a character string normalized to the variant notation.

According to a fourth aspect of the present invention, in the document search system according to any one of the first to third aspects, when the client side reversely displays a character string hit in the search, the client uses a different notation regularity. It is possible to eliminate the need for conversion work.

According to the fifth aspect of the present invention, in a search that absorbs fluctuations in a different notation character string, it is possible to perform inverted display before normalization of the different notation by pattern matching on the output side of the search result.

According to a sixth aspect of the present invention, in the program according to the fifth aspect, text can be extracted from the registration request document and an index can be created for the full text which has been normalized in a different notation.

The invention according to claim 7 is the invention according to claim 5 or 6
If the search hits in the program described in
It is possible to output a character string before being normalized in a different notation, as a search result, instead of a character string normalized in a different notation.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of a document search system according to an embodiment of the present invention.

FIG. 2 is a flowchart illustrating a process performed by a document registration request processing unit of the document search system.

FIG. 3 is a flowchart illustrating a process performed by a document search request processing unit of the document search system.

FIG. 4 is a block diagram showing an electrical connection of the document search system.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Document search system 3 Database 4 Document registration request processing means 5 Document search request processing means

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Hiroshi Takekawa 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Company (72) Inventor Tetsuya Ikeda 1-3-6 Nakamagome, Ota-ku, Tokyo Stock Inside Ricoh Company (72) Inventor Takuya Hiraoka 1-3-6 Nakamagome, Ota-ku, Tokyo Stock Company Ricoh Company (72) Katsumi Kanazaki 1-6-6 Nakamagome, Ota-ku, Tokyo Stock Company Ricoh Company F Terms (reference) 5B009 QA12 QA15 RB32 VA02 5B075 ND03 NK02 NK35 NK49 PQ02 PQ22

Claims (7)

[Claims] 1. An index of a document requested to be registered is created, the document is registered in a database, and a plurality of documents in the database are searched based on a given search character string using the index. In the document search system for performing the above, the registration request document is normalized in different notation and the index is created, and the character string before being normalized in different notation is retained in the created index even after the normalized notation. Document registration request processing means, performing the search by normalizing the search character string in a different notation,
A document search system, comprising: a document search request processing unit that outputs a document hit in the search and a character string hit in the search before being normalized in a different notation. 2. The method according to claim 1, wherein the document registration request processing unit extracts a text from the registration request document and creates the index for the all-text normalized normalization. Document search system. 3. The document search request processing unit outputs a character string before being subjected to the abnormal notation normalization in the index, in place of the character string normalized by the abnormal notation for the document hit in the search. The document search system according to claim 1, wherein: 4. A client-server system, wherein the document registration request processing means and the document search request processing means perform a different notation normalization operation on a server, so that the client can convert the character string hit in the search into the character string. 4. The document search system according to claim 1, wherein the document is inverted and displayed on a display without normalization of a different notation. 5. An index of a document requested to be registered is created, the document is registered in a database, and a plurality of documents in the database are searched based on a given search character string using the index. In a program that causes a computer to execute the process of performing the above, an index is created by normalizing the registration request document in a different notation, and the character string before creation of the character string before the different notation normalization after this notation normalization is performed. Document registration request processing to be held therein, and performing the search by normalizing the search character string in a different notation,
A program for causing a computer to execute a document search request process for outputting a document that has been hit in the search and a character string that has been hit in the search before being normalized in a different form. 6. The program according to claim 5, wherein in the document registration request processing, text is extracted from the registration request document and the index is created with respect to the all-text normalized normalization. . 7. The document search request process according to claim 1, further comprising outputting, in the index, a character string that has not been normalized in the index, in place of the normalized character string in the document hit in the search. The program according to claim 5 or 6, wherein