JP2012083929A - File retrieval device and file retrieval program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently retrieve a file in a file retrieval environment where a file that should be available for full-text retrieval and a file that does not need the full-text retrieval exist.SOLUTION: An index 1 creation unit 11 refers to a retrieval target file 43 to create an index 1 file 21. The index 1 file 21 includes plural index 1 records having keywords and the like extracted from a file name and file content of each retrieval target file 43. An index 2 creation unit 12 refers to the retrieval target file 43 to create an index 2 file 22. The index 2 file 22 includes plural index 2 records having meta data (such as a document title) representing the file name and file content of each retrieval target file 43. A retrieval unit 13 retrieves each index 1 record for full-text retrieval, and retrieves each index 2 record when the full-text retrieval is unnecessary.

Description

  The present invention relates to a file search apparatus and a file search program for creating an index file in advance for a search target file and searching for the file by referring to the index file.

  In recent years, digitization of information has been progressing rapidly. In the past, the main point of digitization was to store in the database (Data Base) files that are referred to when a computer system such as a bank account or municipal resident's card executes a core business. It was.

On the other hand, various documents created in daily work in organizations such as companies are now stored as document files on each employee's client PC (Personal Computer) and sent to other client PCs as e-mail attachments. Alternatively, it is stored in the file server as shared information for the entire organization. The document file stored in the file server is referred to by various employees, and is sometimes copied and updated on the client PC.
In this way, a large number of document files are distributed and stored in various computers, so that a large number of duplicate or similar document files exist in the organization. In addition, the storage capacity of each computer continues to increase.

On the other hand, there are many cases in the organization where it is desired to refer to a document file whose storage location is unknown, for example, a document file created once by an already retired employee is required. In such a case, a method of searching a document file by a full-text search using a keyword that should be included in the document file is widely used.
However, if a document file is to be searched from all computers in the organization, it is necessary to permit access from all employees to the client PC of each employee, which is not preferable in terms of security. For this reason, a document file to be shared by the entire organization is generally stored in a predetermined file server.

However, even if document files are stored in some predetermined file servers, every time a document file is needed, if all document files stored in the file server are searched, a file IO (Input Output) It takes a huge amount of time to do so.
Therefore, the file search device disclosed in Patent Document 1 uses a relatively small amount of information (index information) to be used for a search, such as a storage location or a keyword, for an index file. Remember as. That is, in this file search apparatus, by creating an index file, it is possible to acquire index information for a plurality of files with a single file IO at the time of search, and file IO for each file is unnecessary. Can do. As a result, the response time when searching for a file can be shortened, and the load on the file server can be reduced.

JP 2003-162545 A

However, in order to perform a file search by the file search apparatus disclosed in Patent Document 1, it is necessary to create an index file for all files that are to be searched.
The storage capacity of the index information is relatively small compared to the file itself. However, in order to be able to perform a full text search using keywords, it is necessary to analyze the keywords contained in the file and store them in the index information. For this reason, as the number of keywords included in the file increases, a larger capacity is required. Accordingly, as the number of files to be searched increases, the storage capacity of the index file increases.

  Here, for example, when considering a file search in an organization, even if the file is stored in a shared file server, the accessible folder for each department is limited, and the subordinate folder such as the limited folder is under control. In many cases, files are stored and searched. In such a case, a name such as “work report folder” is given to the folder, and a file matching the name is stored in the folder, that is, the file is classified by the folder. In many cases, the method of doing is taken. In such a case, since a necessary file can be searched by manipulating the tree structure of the folder, it is rare that a full-text search using a keyword is required.

On the other hand, files obtained from outside the organization, such as patent documents and technical papers, and files existing on a server outside the organization, such as a WEB server, may be referred to for purposes that were not intended at the time of acquisition. In many cases, full text search using keywords is required.
Considering file search in an organization in this way, there are files that should be capable of full text search and files that do not require full text search. This is not limited to an organization, and the same applies to, for example, a personal file search.

  An object of the present invention is to provide a file search apparatus and a file search program capable of efficiently searching for a file in a file search environment in which there are files that should be capable of full-text search and files that do not require full-text search. Is to provide.

In order to achieve the above object, the file search apparatus of the present invention provides:
A file acquisition means for acquiring a first search target file group consisting of a plurality of search target files and a second search target file group consisting of a plurality of search target files from the storage device;
For each search target file included in the first search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored Index 1 file storage for creating an index 1 record having a specified file path and a plurality of keywords extracted from the contents of each search target file at a predetermined timing, and storing the index 1 file including each index 1 record Means,
For each search target file included in the second search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored are shown. Index 2 file storage for creating an index 2 record having a file path and one or more metadata representing the contents of each search target file at a predetermined timing and storing an index 2 file including each index 2 record Means,
Keyword search means for searching for each index 1 record included in the index 1 file and acquiring an index 1 record satisfying the search condition;
Metadata search means for searching for each index 2 record included in the index 2 file and obtaining an index 2 record satisfying the search condition;
Is provided.

Preferably, the file search device of the present invention is
A virtual classification display name indicating the name of the virtual classification, an index 2 record presence / absence information indicating whether any index 2 record included in the index 2 file may be classified into the virtual classification, and the index 2 file In order to identify the classification condition for classifying the search target file corresponding to each index 2 record included in the virtual classification into the virtual classification, the upper hierarchy presence / absence information indicating the presence / absence of the upper hierarchy virtual classification, and the upper hierarchy virtual classification Virtual classification definition file storage means for storing a virtual classification definition file including a plurality of virtual classification definition records having higher layer identification information;
Based on the index 2 record presence / absence information, the classification condition, the upper hierarchy presence / absence information, and the upper hierarchy identification information of each virtual classification definition record included in the virtual classification definition file stored by the virtual classification definition file storage unit, Virtual classification definition means for creating virtual classification definition information indicating a tree-like hierarchical structure constituted by a classification and a search target file corresponding to each index 2 record;
Outputs the virtual classification definition information created by the virtual classification definition means, the virtual classification display name of each virtual classification constituting the tree-like hierarchical structure indicated by the virtual classification definition information, and the file name of each search target file Virtual classification output means,
Is provided.

Preferably, the file search device of the present invention is
Keyword dictionary file storage means for storing a keyword dictionary file including a plurality of keyword dictionary records having a keyword and keyword identification information for identifying the keyword;
A plurality of keyword related definition records having keyword identification information of a keyword dictionary record included in the keyword dictionary file and keyword identification information of a keyword dictionary record included in one or more of the keyword dictionary files related to the keyword identification information. Keyword-related definition file storage means for storing a keyword-related definition file including:
When searching by the metadata search means, keywords included in the search conditions of the search, keywords and keyword identification information of each keyword dictionary record included in the keyword dictionary file, and included in the keyword related definition file Virtual classification dynamic generation means for generating the virtual classification definition record based on the keyword identification information of each keyword-related definition record and one or more keyword identification information related to the keyword identification information;
Is provided.

Preferably, the file search device of the present invention is
The index 2 record includes index 2 record identification information for identifying the index 2 record,
One or more related definition records having related definition record identification information for identifying the related definition record, a display name indicating the name of the related definition, and a search condition acquisition condition for acquiring the search condition from the index 2 record. An associated definition file storage means for storing an associated definition file including:
Relation definition information output means for outputting relation definition record identification information and display name of each relation definition record included in the relation definition file storage means;
Receiving means for receiving related definition record identification information selected from each related definition record identification information output by the related definition information output means, and index 2 record identification information;
Search condition acquisition means for acquiring a search condition from an index 2 record including the index 2 record identification information received by the reception means based on a search condition acquisition condition of each related definition record included in the related definition file storage means; ,
Each index 2 record included in the index 2 file is searched, an index 2 record satisfying the search condition acquired by the search condition acquisition unit is acquired, and the file name and file path included in the acquired index 2 record Related search means for outputting
Is provided.

The file search program of the present invention is
Computer
A file acquisition means for acquiring a first search target file group consisting of a plurality of search target files and a second search target file group consisting of a plurality of search target files from the storage device;
For each search target file included in the first search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored Index 1 file storage for creating an index 1 record having a specified file path and a plurality of keywords extracted from the contents of each search target file at a predetermined timing, and storing the index 1 file including each index 1 record Means,
For each search target file included in the second search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored are shown. Index 2 file storage for creating an index 2 record having a file path and one or more metadata representing the contents of each search target file at a predetermined timing and storing an index 2 file including each index 2 record Means,
Keyword search means for searching for each index 1 record included in the index 1 file and acquiring an index 1 record satisfying the search condition;
Metadata search means for searching for each index 2 record included in the index 2 file and obtaining an index 2 record satisfying the search condition;
And make it work.

  ADVANTAGE OF THE INVENTION According to this invention, a file can be searched efficiently in the file search environment where the file which should enable full text search, and the file which does not require full text search exist.

It is a figure which shows an example of a structure of the file search system which concerns on 1st Embodiment. It is a figure which shows an example of a data structure of an index 1 file. It is a figure which shows an example of a data structure of an index 2 file. It is a figure which shows an example of a data structure of the system metadata contained in an index 2 file. It is a figure which shows an example of a data structure of the standard metadata contained in an index 2 file. It is a figure which shows an example of the data structure of a virtual classification definition file. It is a figure which shows an example of a data structure of a related definition file. It is a flowchart which shows an example of operation | movement of an index 1 preparation part. It is a flowchart which shows an example of operation | movement of the index 2 preparation part. It is a flowchart which shows an example of operation | movement of a search request | requirement part. It is a flowchart which shows an example of operation | movement of a search part. It is a flowchart which shows an example of operation | movement of an index 1 search part. It is the figure which illustrated the login screen. It is the figure which illustrated the search request screen. It is the figure which illustrated the search request screen. It is a flowchart which shows an example of operation | movement of the search part etc. in a compound search. It is the figure which illustrated the display contents of the search result on a search request screen. It is the figure which illustrated the display contents of the search result on a search request screen. It is a figure which shows an example of a structure of the file search system which concerns on 2nd Embodiment. It is a figure which shows an example of a data structure of a keyword dictionary file. It is a figure which shows an example of a data structure of a keyword related definition file. It is a flowchart which shows an example of operation | movement of the search part etc. in virtual classification dynamic generation. It is the figure which illustrated the search request screen.

Hereinafter, a file search system according to a first embodiment of the present invention will be described with reference to the drawings.
In the present specification and claims, unless otherwise specified, the “file” includes not only a file such as a document file and an image file, but also the entire database and each record of the database. This means electronic data that can be subject to viewing, mail transmission / reception, copying to an external storage medium, and the like.

<Configuration and function of the entire file search system>
FIG. 1 shows an example of the configuration of a file search system according to the first embodiment of the present invention.
The file search system according to the first embodiment includes a file search server 1, a file server 4, and a client 3. The file search server 1, the file server 4, and the client 3 are communicably connected via a wired or wireless communication line 9 such as a LAN (Local Area Network).
Here, one file search server 1, one file server 4, and one client 3 are illustrated, but two or more each may exist. The file search server 1, the file server 4, and the client 3 do not have to be different devices. For example, the functions of the file search server 1, the file server 4, and the client 3 are realized by a single device. Is also possible.
The communication line 9 is not limited to a LAN, and can be, for example, a WAN (Wide Area Network), the Internet, or a combination thereof.

  As will be described in detail later, the file search server 1 collects file names and the like stored in the file server 4 to create and store an index file. When the client 3 transmits a file search request (hereinafter sometimes abbreviated as “search request”) to the file search server 1, the file search server 1 performs a file search with reference to the index file, and the client 3 The file search result (hereinafter may be abbreviated as “search result”) is sent to.

<Configuration and function of client 3>
The client 3 is a computer and includes an input device 32, a display device 33, a main storage device and a storage device (not shown).
The input device 32 includes a keyboard, a mouse, and the like. The operator of the client 3 can instruct a process to be executed by the client 3 by operating the input device 32. That is, the input device 32 functions as input means for the client 3.
The display device 33 includes a liquid crystal display, a printer, and the like. The display device 33 displays or prints the result of processing executed by the client 3. That is, the display device 33 functions as a display unit and an output unit of the client 3.
The main memory is RAM (Random
Access Memory). The storage device includes a magnetic disk or the like, and is built in the client 3 or externally connected. The main storage device and the storage device function as storage means of the client 3.

Although not shown, the client 3 includes a CPU (Central Processing Unit). A program such as a search request program is stored in the main storage device. The function of the search request unit 31 and the like is realized by the CPU executing the instruction code.
Further, when executing the instruction code of an application program such as a search request program, the CPU may also execute the instruction code of a program such as an OS (Operating System). Since the technique related to program execution as described above is well known, in the following description and drawings, it is as if there is hardware that executes processing realized by various programs, such as the search request unit 31. It describes. It should be noted that each unit (for example, the search request unit 31) can actually be configured by hardware.

The search request unit 31 receives a file search condition and a file search instruction (for example, an instruction to search for a file whose file name is “work report”) input by the operator of the client 3 using the input device 32. The search condition formula is created, and a search request including the search condition formula is transmitted to the file search server 1.
Further, the search result transmitted from the file search server 1 to the client 3 is received and displayed on the display device 33. The search request program for realizing the search request unit 31 may be a unique program according to the first embodiment, or may be a WEB browser program, for example. When a WEB browser is used as the search request unit 31, the search unit 13 of the file search server 1 to be described later may be a WEB application, for example.

<Configuration / Function of File Server 4>
The file server 4 is a computer and includes a storage device 42.
The storage device 42 includes a magnetic disk or the like. The storage device 42 is built in the file server 4 or connected externally. In FIG. 1, an example is shown as if one storage device 42 is connected to the file server 4, but actually, two or more storage devices 42 may rather be connected to the file server 4. In many cases, two or more storage devices 42 are connected to two or more file servers 4 in a switchable manner.
Various files are stored in the storage device 42. In FIG. 1, a file to be indexed by the file search server 1 described later is described as a “search target file 43”. That is, a specific file called the search target file 43 does not exist, and the search target file 43 may be, for example, all the files stored in the storage device 42 or all the files in the specific folder Some of the files stored in the storage device 42 may be used.
In the search target file 43, access authority representing reference / update / execution authority for the owner, group, and other users can be set.

  Although not shown, the file server 4 includes a CPU and a main storage device including a RAM and the like. A program such as a file management program is stored in the main storage device. Functions of the file management unit 41 and the like are realized by the CPU executing the instruction code. The file management unit 41 manages the storage location and the like of files stored in the storage device 42 including the search target file 43. In addition, the file management unit 41 receives the storage location of the file from the file search server 1, reads the search target file 43 stored in the storage location, and transmits the storage content of the file to the file search server 1. It also has a function.

<Configuration / Function of File Search Server 1>
The file search server 1 is a computer and includes a storage device 2 and a main storage device (not shown). The storage device 2 includes a magnetic disk or the like. The storage device 2 is built in the file server 1 or connected externally. In FIG. 1, an example is shown as if one storage device 2 is connected to the file search server 1, but there may be two or more. The main storage device is composed of a RAM or the like. The storage device 2 and the main storage device function as storage means of the file server 1.
The storage device 2 stores an index 1 file 21, an index 2 file 22, a virtual classification definition file 23, a related definition file 24, a keyword dictionary file 25, and a keyword related definition file 26. The contents stored in these files will be described together with the functional description of the file search server 1 hereinafter.

The file search server 1 includes a CPU (not shown). A program such as a file search program is stored in the main storage device of the file search server 1. When the CPU executes the instruction code, each function of the index 1 creation unit 11, the index 2 creation unit 12, the search unit 13, and the index 1 search unit 14 is realized.
The index 1 creation unit 11 creates the index 1 file 21 with reference to the search target file 43 at a predetermined timing such as the same time every day. As will be described later, the index 1 file 21 stores a file name, keywords extracted from the file contents, and the like.
Which file is the search target file 43 is not shown, for example, but one or more file paths (such as “/ etc / usr1 /”) of the storage device 42 are stored in advance in a predetermined file in the storage device 2. In addition, the file under the file path can be set as the search target file 43. When there are two or more storage devices 42, etc., the storage device including the storage device on which the file path exists may be stored. By creating the index 1 file 21 as described above, it becomes possible to search for a file by referring to the index 1 file 21, and it is not necessary to refer to the search target file 43 every time it is searched. Processing time can be shortened.

When searching with reference to the index 1 file 21 created as described above, the search target file 43 may be deleted or changed after the index 1 file 21 is created. In such a case, when the index 1 file 21 is searched, a search result different from that when the search target file 43 is searched is obtained. For example, if an index 1 file 21 is created for a file whose file name is “work report” and then the file is deleted, when the index 1 file 21 is referenced, there is a file whose file name is “work report”. The search target file 43 does not exist.
Therefore, as described above, the index 1 creation unit 11 can perform processing at a predetermined timing such as the same time every day. By doing so, it is possible to prevent the index 1 file 21 from being periodically updated and greatly deviating from the search target file 43.

Then, by shortening the interval of the timing at which the index 1 creation unit 11 performs processing (for example, by performing processing every hour), the divergence between the index 1 file 21 and the search target file 43 is further increased. Can be small. However, shortening the interval at which the index 1 creation unit 11 performs processing means that the interval at which IO is generated for all files of the search target file 43 is also shortened, so the performance of the file server 4 is also taken into consideration. In the above, it is necessary to determine the timing interval at which the index 1 creation unit 11 performs processing.
For example, when the file server 4 runs a program that constantly monitors the CPU usage rate, the IO frequency at a certain time, etc., and the CPU usage rate, the IO frequency at a certain time, etc. become below a predetermined value, that fact Is also transmitted to the index 1 creation unit 11 and the index 1 creation unit 11 starts processing. Alternatively, in the file server 4, a program that constantly monitors the IO in the search target file 43 is run, and whenever the search target file 43 is updated, the index 1 creation unit 11 is notified of this, and the index 1 The index information for the file in the file 21 may be updated.

  Similarly to the index 1 creation unit 11, the index 2 creation unit 12 also creates the index 2 file 22 by referring to the search target file 43 at a predetermined timing such as the same time every day. In the first embodiment, as will be described later, for example, the index 2 file 22 stores the title of the document stored in the file determined from the contents of the file. That is, the stored contents of the index 1 file 21 and the stored contents of the index 2 file 22 are not completely the same, even if some of them overlap.

  As with the index 1 creation unit 11, which file is the search target file 43 is set to one or more file paths (such as “/ etc / usr1 /”) of the storage device 42 in advance in a predetermined file in the storage device 2. The file under the file path can be stored as the search target file 43. The search target file 43 of the index 1 creation unit 11 and the search target file 43 of the index 2 creation unit 12 may be the same, a part of them may overlap, or they may be completely different.

  When searching with reference to the index 2 file 22 created as described above, the stored contents of the index 2 file 22 and the search target file 43 are searched in the same manner as when searching with reference to the index 1 file 21. There arises a problem that the contents may deviate. This problem can be solved by shortening the interval of the timing at which the index 2 creation unit 12 performs processing, as with the index 1 creation unit 11.

  Furthermore, when the creation timings of the index 1 file 21 and the index 2 file 22 are different, the contents of the index 1 file 21 and the index 2 file 22 may be different from each other. One solution to this problem is to match the index file creation processing start timings of the index 1 creation unit 11 and the index 2 creation unit 12. Specifically, for example, immediately before the index 2 creation unit 12 starts the index file creation process, an index creation process start request is transmitted to the index 1 creation unit 11, and when the index 1 creation unit 11 receives the request, the index 1 creation unit 11 The creation process may be started.

Of course, even if the index file creation processing start timings coincide, the timings of referring to the search target files 43 do not necessarily coincide, but even if the contents of the index 1 file 21 and the index 2 file 22 are slightly different from each other This does not mean that there will be a major hindrance to use for search purposes. Therefore, if the index file creation processing start timings are matched, there is no practical problem in most cases.
However, there may be a case where it is necessary to completely prevent the difference between the contents of the index 1 file 21 and the index 2 file 22. In such a case, for example, in the file server 4, a program for constantly monitoring the IO in the search target file 43 is run, and each time the search target file 43 is updated, the index 1 file 21 and the index 2. Information about the file in the file 22 may be updated.

  The search unit 13 is activated when the file search server 1 receives a search request from the client 3 and executes a file search. More specifically, one or both of the index 1 file 21 and the index 2 file 22 are referred to determine whether there is a file that matches the search condition expression of the search request. The search result including the file name and the like is transmitted to the client 3.

Here, when it is necessary to refer to the index 1 file 21, the search unit 13 passes the search condition expression to the index 1 search unit 14, and the index 1 search unit 14 refers to the index 1 file 21 and receives it. It is determined whether there is a file that matches the search condition expression. Since the index 1 file 21 stores keywords extracted from the file name and file contents, the search by the index 1 search unit 14 can complement a part of the search function of the search unit 13.
As will be described later, in the first embodiment, when the operator of the client 3 requests a so-called full-text search, the index 1 search unit 14 performs a search. By separating the full-text search function from the function of the search unit 13 itself, various existing full-text search programs can be used as programs for realizing the index 1 search unit 14, for example, search requests A full text search is performed by the index 1 search unit 14 that matches a search purpose, such as whether a search for a document in the social science field is requested or whether a search for a document in the natural science field is required. It is also possible to do.

  The search unit 13 classifies the search target file 43 according to metadata search keywords such as document titles stored in the index 2 file 22 and creates a tree-like hierarchy (hereinafter referred to as “virtual classification function”). In other words, it has a function of classifying the index 2 record 220 (to be described later with reference to FIG. 3), and when performing virtual classification, a virtual classification definition file 23 that defines classification conditions and the like. Refer to

The search unit 13 has a function (hereinafter referred to as “virtual classification dynamic generation function”) that dynamically generates a virtual classification definition using a keyword related to a keyword for metadata search. By using the virtual classification function with reference to the virtual classification definition generated by the virtual classification dynamic generation function, classification using keywords related to the metadata search keyword can be performed.
When the virtual classification dynamic generation is performed, the keyword dictionary file 25 for registering keywords and the keyword association definition file 26 that defines the association of keywords registered in the keyword dictionary file 25 are referred to. Of course, there are programs for creating / updating the keyword dictionary file 25 and the keyword related definition file 26. However, since they are not directly related to the present invention, the keyword dictionary file 25 and the keyword related definition file 26 will be described in the following description. It will be explained on the assumption that it has been created.

The search unit 13 refers to the storage position of the search target file 43 stored in the index 2 file 22 on the storage device 42 and creates a tree-like hierarchy (hereinafter referred to as “physical hierarchy creation function”). .)have.
Further, the search unit 13 has a function of searching for a file related to the search result (hereinafter referred to as “related search function”), and when performing a related search, a related definition that defines a related search condition or the like. Refer to the file 24. Of course, there is a program for creating / updating the related definition file 24, but since it is not directly related to the present invention, the following description will be made on the assumption that the related definition file 24 has been created.

<Configuration and function of each file>
FIG. 2 shows an example of the data structure of the index 1 file 21.
The index 1 file 21 includes index 1 records 210 corresponding to the search target files 43. That is, each index 1 record 210 has a one-to-one correspondence with the search target file 43 when the index 1 record 210 is created.
The index 1 record 210 includes data items of a file name 211, a file path 212, an access authority 213, and a keyword 214.

In the file name 211, the file name of the corresponding search target file 43, for example, “work report 1.doc” is set.
The file path 212 includes an absolute path of the corresponding search target file 43, for example, “// etc / usr1 / work report 1.doc” (work report 1 in the “usr1” folder in the “etc” folder immediately below the root). .doc ") is set. Note that the storage location of the search target file 43 on the storage device 42 can be identified by the file path 212, but when a plurality of storage devices 42 are connected to the file server 4, The logical volume name and the like may be set as a part of the file path 212 or as a data item different from the file path 212. It is also possible to identify the storage location of the search target file 43 by information other than the absolute path of the file, for example, a relative path based on a predetermined file or a logical block number in which the file is stored. It is. Such data items may be provided in the index 1 record 210 in place of or in addition to the file path 212.

In the access authority 213, the access authority set for the corresponding search target file 43 is set. More specifically, for example, the access authority (for example, UNIX (registered trademark), etc., which is given by the file management unit 41 of the file server 4 and stored as file attribute information, etc., owner, group, etc. A three-digit numerical value (such as 777) representing the reference / update / execution authority for the user is set.
In the keyword 214, a keyword extracted from the content of the corresponding search target file 43 is set. For example, when the search target file 43 includes a plurality of character strings “site”, one or more keywords are extracted by various syntax analysis methods such as extracting “site”, and the keyword 214 is extracted. Can be set. In general, a large number of phrases and the like are stored in the keyword 214, and most of the size of the index 1 record 210 is used for the keyword 214.
In addition to the above, other file attribute information such as a file creator can also be a data item of the index 1 record 210.
The index 1 record 210 (index information) created as described above is referred to by the index 1 search unit 14.

FIG. 3 shows an example of the data structure of the index 2 file 22.
The index 2 file 22 includes index 2 records 220 corresponding to the search target files 43. That is, each index 2 record 220 has a one-to-one correspondence with the search target file 43 when the index 2 record 220 is created.
The index 2 record 220 includes system metadata 221, standard metadata 222, and user-defined metadata 223.
The system metadata 221 is set by the index 2 creation unit 12, and the user cannot change the setting contents directly.
On the other hand, metadata representing the contents of the file is set in the standard metadata 222 by the index 2 creation unit 12. Although not shown in FIG. 1, the user can directly change the setting contents of the standard metadata 222 by causing the CPU of the file search server 1 to execute a metadata change program.
User-defined metadata 223 is a data item in which a user defines a data structure and sets / changes metadata representing the contents of a file. The index 2 creation unit 12 does not set the user-defined metadata 223.

FIG. 4 is a diagram illustrating an example of the data configuration of the system metadata 221.
The system metadata 221 includes a file ID 221a, a file name 221b, and a file path 221c.
In the file ID 221a, an ID (Identifier) capable of uniquely identifying the corresponding search target file 43 is set. Specifically, for example, each time a new index 2 record 220 of the search target file 43 is created, a serial number starting from 1 may be set.
Similarly to the file name 211 of the index 1 record 210, the file name of the corresponding search target file 43, for example, “work report 1.doc” is set in the file name 221b.
Similar to the file path 212 of the index 1 record 210, for example, the absolute path (“// etc / usr1 / work report 1.doc” or the like) of the corresponding search target file 43 is set in the file path 221c.
In addition to the above, other file attribute information such as file creator and access authority can also be used as data items of the system metadata 221.

FIG. 5 is a diagram illustrating an example of a data configuration of the standard metadata 222.
The standard metadata 222 includes a title 222a, a document description date 222b, and a security rank 222c.
In the title 222a, the title of the document or the like stored in the corresponding search target file 43 is set as “work report”. Specifically, for example, the index 2 creation unit 12 constructs a display image when the file is printed, and sets a character string to be printed with a character size larger than other characters on the top of the first page as a title of a document or the like. Is set as the title 222a.
In the document description date 222b, the date when the document stored in the corresponding search target file 43 is described is set as “August 5, 2009”. Specifically, the index 2 creation unit 12 includes, for example, “year”, “month”, “day”, and “creation” among the character strings printed on the top of the first page when the file is printed. The character string that seems to be the creation date is regarded as the creation date of the document, and is set as the document description date 222b.

In the security rank 222c, the confidentiality of the document or the like stored in the corresponding search target file 43 is set as “top secret” or “secret”. Specifically, it is considered that the index 2 creation unit 12 indicates that it is necessary to keep confidential information such as “handling precautions” and “duplication prohibited” in the character string printed in the file. The character string is extracted, the confidentiality is determined based on the content, number, etc. of the extracted character string, and set to the security rank 222c.
In addition to the above, other information that can be determined from a display image when the search target file 43 is printed, such as a document storage period, can also be used as data items of the standard metadata 222.
The index 2 record 220 (index information) created as described above is referred to by the search unit 13 as described above.

FIG. 6 shows an example of the data configuration of the virtual classification definition file 23.
The virtual classification definition file 23 is composed of one or more virtual classification definition records 230. The virtual classification definition record 230 includes data items of a virtual classification ID 231, a display name 232, a condition 233, and a higher virtual classification ID 234.

In the virtual classification ID 231, values that can uniquely identify the virtual classification definition record 230 are set as “1”, “2”, and “3”.
In the display name 232, the name of the virtual classification is set as “title”, “work report”, and the like.
The condition 233 is set such that the classification condition based on the virtual classification is “no condition” and “the title 222a includes“ work report ”(character string).” The classification condition is “no condition”. If it is set, it means that there is no index 2 record 220 classified in the virtual classification, and if any condition is set in the classification condition, a record that satisfies the condition in the index 2 record 220 Therefore, one index 2 record 220 may be classified into two or more virtual classifications, or may not be classified into any virtual classification.
In the upper virtual classification ID 234, values that can uniquely identify the upper virtual classification definition record 230 of the virtual classification definition record 230 are “0 (no upper)”, “1”, “2”, “3”, and the like. Set to

  As will be described later with reference to FIGS. 14 and 15, the virtual classification definition record 230 is referred to by the search unit 13, and the display name 232 and the index 2 record 220 satisfying the condition 233 related to the display name 232 are stored. The file name 221b is hierarchically displayed on the display device 33 in a tree shape. Therefore, one index 2 record 220 may be displayed at two or more places in the tree, or may not be displayed anywhere in the tree.

FIG. 7 shows an example of the data structure of the related definition file 24.
The related definition file 24 is composed of one or more related definition records 240. The related definition record 240 includes data items of a related definition ID 241, a display name 242 and a condition 243.
In the related definition ID 241, values that can uniquely identify the related definition record 240 are set as “1”, “2”, and “3”.
In the display name 242, the name of the relevant definition is set as “title”.
In the condition 243, the classification condition based on the virtual classification is set such that “title 222a is equal to the search result”. The condition 243 includes various conditions such as “a copy is added before the file name of the search file” and “a number is added after the file name of the search file”. Various conditions may be settable.
The related search using the related definition file 24 will be described in detail later with reference to FIGS. 17 and 18.

20 and 21 show examples of data structures of the keyword dictionary file 25 and the keyword related definition file 26, respectively.
The virtual classification definition record 230 described above may be created in advance by the user, but can be dynamically created using the keyword dictionary file 25 and the keyword related definition file 26 when performing a metadata search. It is.
The keyword dictionary file 25 is composed of one or more keyword dictionary records 250. The keyword dictionary record 250 includes data items of a keyword ID 251, a language code 252, and a keyword 253.
In the keyword ID 251, values that can uniquely identify the keyword are set as “1”, “2”, “3”.
In the language code 252, a three-letter language code representing the language type of the keyword 253 is set as “jpn” or “eng”. The language codes “jpn” and “eng” correspond to Japanese and English, respectively.
In the keyword 253, the keyword is set as “specification”.
The keyword dictionary record 250 can be uniquely identified by the keyword ID 251 and the language code 252. Further, a keyword dictionary record 250 having the same keyword 253 cannot be set in the keyword dictionary file 25.
For example, keyword dictionary record 250 in which “1”, “jpn”, and “specification” are set in keyword ID 251, language code 252, and keyword 253, and “1”, “eng”, and “specifications” are set, respectively. If there are two recorded keyword dictionary records 250, the keywords of these two records are the same keyword although the languages are different.

The keyword related definition file 26 includes one or more keyword related definition records 260. The keyword related definition record 260 includes a keyword ID 261, a related keyword ID 262, a maximum keyword length 263, and a language condition 264.
In the keyword ID 261, the value of the keyword ID 251 of the keyword dictionary record 250 registered in the keyword dictionary file 25 is set.
In the related keyword ID 262, one or more values of the keyword ID 251 of the keyword dictionary record 250 registered in the keyword dictionary file 25 are set as in the keyword ID 261. The keyword related definition record 260 is a keyword dictionary record having a keyword ID 251 whose value is the same as that of the keyword set in the keyword 253 included in the keyword dictionary record 250 of the keyword ID 251 having the same value as the keyword ID 261. It is designated that the keyword set in the keyword 253 included in 250 is related.
When two or more keyword IDs are set in the related keyword ID 262, they are set by being separated by “, (comma)” such as “1, 2”.

In the maximum keyword length 263, the limit on the maximum number of characters of the keyword when the keyword of the related keyword ID 262 is acquired from the keyword dictionary file 25 is set to “0 (no limit)” or “10”.
In the language condition 264, the value of whether or not to limit the language type of the keyword when acquiring the related keyword from the keyword dictionary file 25 to the same as the language type of the search keyword is “0 (only the same language)”, “ 1 (not limited) ”is set.

For example, it is assumed that the keyword ID 251, the language code 252, and the keyword 253 of the keyword dictionary file 25 have three keyword dictionary records 250 set as follows.
(1) “1”, “jpn”, “specifications”
(2) “1”, “eng”, “specifications”
(3) “2”, “jpn”, “design document”
Then, “2”, “1”, “3”, “0 (same language only)” are set in the keyword ID 261, the related keyword ID 262, the maximum keyword length 263, and the language condition 264 of the keyword related definition file 260, respectively. It is assumed that a keyword related definition record 260 exists. In this case, the keyword “specification” is related to the keyword “design document”. However, the keyword “specifications” is not regarded as a related keyword because the language condition 264 of the keyword related definition record 260 is “0 (only the same language)”.
The dynamic creation (virtual classification dynamic generation function) of the virtual classification definition record 230 using the keyword dictionary file 25 and the keyword-related definition file 26 will be described later with reference to FIGS. 22 and 23.
The description of the configuration and functions of the file search system according to the first embodiment has been completed, and the operation of the file search system according to the first embodiment will be described below with reference to the flowcharts of the processing of each unit.

<Operation of each program>
FIG. 8 is a flowchart showing an example of the operation of the index 1 creation unit 11.
When the process is started, the index 1 creation unit 11 creates the index 1 record 210 of the search target file 43 (S801).
Specifically, for example, the file management unit 41 reads and transmits a file included in a file path (for example, “/ etc / usr1 /”) stored in advance in a predetermined file in the storage device 2. To request. When a file is received from the file management unit 41, is there an index 1 record 210 in which the file path of the acquired file (for example, “/ etc / usr1 / work report 1.doc”) is set in the file path 212? judge. If not, the keyword is extracted from the acquired file, and the file name, file path, access authority, and extracted keyword of the file are set in the file name 211, file path 212, access authority 213, and keyword 214, respectively. Then, an index 1 record 210 is added. On the other hand, if it exists, the access authority 213 and the keyword 214 of the index 1 record 210 (hereinafter referred to as “update target record”) are updated.

When the above processing is executed for all files under the preset file path, and there is an index 1 record 210 that has not become an update target record in records other than the newly created index 1 record 210, The index 1 record 210 is deleted.
The method for creating the index 1 record 210 is not limited to the method described above. For example, after the processing starts, the index 1 file 21 is deleted, and the file management unit 41 reads and transmits the file included in the preset file path (for example, “/ etc / usr1 /”). An index 1 record 210 may be created for the requested and received file.

  As described above, in the first embodiment, each time the index 1 creation unit 11 performs processing, an index 1 record 210 corresponding to the search target file 43 at the time of processing is created.

FIG. 9 is a flowchart illustrating an example of the operation of the index 2 creation unit 12.
When the process is started, the index 2 creation unit 12 creates the index 2 record 220 of the search target file 43 (S901). Since the specific processing content is the same as that of the index 1 creation unit 11, only different points will be described below.
First, as described above, the search target file 43 of the index 2 creation unit 12 does not have to be the same as the search target file 43 of the index 1 creation unit 11. For example, all the files stored in the storage device 42 are set as the search target files 43 of the index 1 creation unit 11, and a part of the files stored in the storage device 42 (for example, an operator of the client 3 normally refers to it). File only) can be the search target file 43 of the index 2 creation unit 12. In this way, in the search processing described later, normally, only the index 2 file 22 is searched to reduce the number of files displayed as search results, while the index 1 file 21 is searched as necessary. As a result, files that are not normally referenced can be displayed as search results.

Conversely, a part of the file stored in the storage device 42 (for example, a document file in which a phrase is used relatively strictly, such as a paper or a judgment sentence) is used as the search target file 43 of the index 1 creation unit 11 and stored. All the files stored in the device 42 can be set as the search target files 43 of the index 2 creation unit 12.
By doing so, the possibility that a phrase or the like that is used while the definition is ambiguous is extracted as the keyword 214 of the index 1 record 210 is reduced. Even if the word or the like matches the search keyword and the file is displayed as a search result, there is a relatively high possibility that the file is not actually the desired file. For this reason, the capacity of the index 1 file 21 can be kept relatively small, and when a full-text search using a keyword is performed in the search processing described later, the possibility of including a file to be obtained as a search result can be increased. .

In addition, unnecessary updating of the index 2 record 220 can be avoided by performing the following.
For example, an update date is provided as a data item of the index 2 record 220, and when the index 2 record 220 is created / updated, the year / month / day / hour / minute / second in which the process is performed is set. When the index 2 creation unit 12 tries to update the index 2 record 220, the update date and time of the index 2 record 220 and the update date and time of the search target file 43 (generally, the file management unit 41 is one of the attribute information of the file). If the update date / time of the search target file 43 is newer, the contents of the search target file 43 after the creation of the index 2 record 220 may have changed. Update target.
Further, when the access authority is provided as the data item of the index 2 record 220, when the index 2 creation unit 12 attempts to update the index 2 record 220, the access authority of the index 2 record 220 and the search target file 43 are updated. The access authority is compared, and if they are different, they can be updated.

  Furthermore, when updating the index 2 record 220, the index 2 creation unit 12 determines whether the setting content of the standard metadata 222 has been directly changed using the metadata change program described above, and has been changed directly. In this case, the standard metadata 222 is not updated. In order to do this, for example, “direct change presence / absence” is provided as the data item of the standard metadata 222, and when the index 2 record 220 is created, the data item “direct change presence / absence” is set to “no direct change”. When it is set and changed directly using the metadata change program, “directly changed” may be set in the data item “directly changed”. The index 2 creation unit 12 does not update the user-defined metadata 223.

As a data item of the standard metadata 222, “file update presence / absence after direct change” may be further provided. When the index 2 creation unit 12 creates the index 2 record 220, “no update” is set in the data item “whether or not the file is updated after direct change”. Then, when the index 2 creation unit 12 updates the index 2 record 220, it is determined whether or not the index 2 record 220 is directly changed using the metadata change program, and the index 2 record 220 is directly changed and corresponds. If the contents of the search target file 43 have changed, “updated” is set in the data item “presence / absence of file update after direct change”.
In this way, when the operator of the client 3 refers to the index 2 record 220 using the metadata change program or displays the file search result as described later, the standard metadata 222 is stored. After the direct change, it is possible to notify that the content of the file displayed on the display device 33 has been updated, and the operator of the client 3 needs to directly change the standard metadata 222 again. Can be determined.
As described above, in the first embodiment, each time the index 2 creation unit 12 performs processing, an index 2 record 220 corresponding to the search target file 43 at the time of processing is created.

FIG. 10 is a flowchart illustrating an example of the operation of the search request unit 31.
The search request unit 31 is activated by the operator of the client 3 using the input device 32.
When activated, the search request unit 31 performs a login process (S1001). Specifically, when the login screen as shown in FIG. 13 is displayed on the display device 33, the operator of the client 3 inputs the user ID and password using the input device 32, and presses the “execute” button. It is determined whether or not the input user ID and password are appropriate, and the user ID access authority for the search target file 43 is determined. In addition, since the above login process itself is a well-known technique, further description is abbreviate | omitted.

  When it is determined that the input user ID and password are appropriate, the search request unit 31 displays a search request screen on the display device 33 (S1002). FIG. 14 shows a display example of the search request screen (D1401). As shown in FIG. 14, the search request screen includes a virtual classification display unit (D1402), a physical folder display unit (D1403), a search condition unit (D1404), a search result unit (D1405), “search”, “metadata edit”. ”And“ End ”buttons.

  When displaying the search request screen (D1401), the search request unit 31 displays the virtual classification display unit (D1402) using the virtual classification function of the search unit 13. Specifically, the search request unit 31 requests the search unit 13 to transmit the initial display content of the virtual classification, and the search unit 13 sets “0 ( The display name 232 of the record in which “no higher order” is set (that is, the highest virtual classification definition record 230) is transmitted to the search request unit 31, and the search request unit 31 uses the received display name as the virtual classification display unit. (D1402). In addition, the search request unit 31 displays a figure in which “+” is squared before each display name 232. As will be described later, the operator of the client 3 can display other virtual classifications and files included in the virtual classification by performing an operation such as clicking the figure with a mouse.

For example, a virtual classification definition record 230 in which “1”, “title”, “no condition”, and “0 (no upper)” are set in the virtual classification ID 231, the display name 232, the condition 233, and the upper virtual classification ID 234, respectively. , Suppose that there is a virtual classification definition record 230 in which “2”, “work report”, “title 222a contains“ work report ””, and “1” are set. When a figure surrounded by a square “+” displayed in front of the title of D1402) is clicked, a folder named “Work Report” appears in the folder named “Title” as shown in D1402 of FIG. A tree is displayed as if it were included.
Furthermore, there are four index 2 records 220 in which “work report” is included in the title 222a, and each file name 221b is “work report 1.doc”, “work report 2.doc”, “work report”. 3. In the case of “doc” and “report.doc”, when a figure surrounded by a square “+” displayed before the work report is clicked, as shown in the virtual classification display section (D1402) of FIG. Whether the folder named “Work Report” contains four files: “Work Report 1.doc”, “Work Report 2.doc”, “Work Report 3.doc”, and “Report.doc”. A tree display like this is made.

Further, when displaying the search request screen (D1401), the search request unit 31 displays the physical folder display unit (D1403) using the physical layer creation function of the search unit 13. Specifically, the search request unit 31 requests the search unit 13 to transmit the initial display contents of the physical folder. The search unit 13 refers to the file path 221 c in the system metadata 221 of the index 2 record 220, creates a tree-like hierarchy of folders, and transmits the tree-like hierarchy to the search request unit 31. The search request unit 31 displays the name of the folder included in the received tree-like hierarchy on the physical folder display unit (D1403).
Further, the search request unit 31 displays a figure in which “+” is enclosed in a square before the name of each folder. The operator of the client 3 can display other folders and files included in the folder by performing an operation such as clicking the figure with a mouse.

The display contents of the search condition part (D1404) and the search result part (D1405) are as shown in FIG. 14, and no search result is displayed in the search result part (D1405).
The operator of the client 3 uses the input device 32 to input each item in the search condition part (D1404). The item entered in the search condition part (D1404) is a search condition. For example, when “site” is entered in the full text and “work report” is entered in the title, “site” is set in the keyword 214 of the index 1 record 210. In addition, a file in which “work report” is set in the title 222a of the index 2 record 220 is searched, and the search result is scroll-displayed in the search result section (D1405).

For each item of the search condition part (D1404), the convenience of search can be improved by enabling various input methods. For example, it may be possible to input logical expressions such as “not on-site” and “on-site and work” in the whole text. Note that the input items of the search condition part (D1404) are not limited to the illustrated items, and may be determined according to the data items of the index 1 file 21 and the index 2 file 22, such as access authority and security rank.
After the search request screen is displayed (S1002), the search request unit 31 waits for the search button, metadata edit button, or end button to be pressed (S1003, S1004). When the search button is pressed, that is, when a search request is detected (S1003: YES), search processing (S1005, S1006, S1007) is performed. When the end button is pressed, that is, when an end request is detected (S1004: YES), the process ends.

  Although not shown in FIG. 10, the search request unit 31 performs metadata editing processing when the metadata editing button is pressed. Specifically, the search request unit 31 requests the operator of the client 3 to specify a file to be edited, and the standard metadata 222 and user-defined metadata 223 of the specified file at that time Display the settings. Then, the search request unit 31 changes the setting contents of the standard metadata 222 and the user-defined metadata 223 with the contents input by the operator of the client 3. At this time, as described above, when “updated” is set in the file update presence / absence after direct modification of the standard metadata 222, a message to that effect may be displayed on the display device 33.

When the search request unit 31 detects the search request, the search request unit 31 transmits the input content (search condition) of the search condition unit (D1404) to the search unit 13 (S1005). For example, if “on-site” is entered for the full text and “work report” is entered for the title, a search condition expression such as “full text = site title = work report” is created, and the user ID entered on the login screen and the search section 13 to send. Here, the search condition expression is an expression interpreted by the search unit 13 and may be created according to a syntax rule or the like that can be interpreted by the search unit 13.
After transmitting the search condition formula to the search unit 13, the search request unit 31 waits until a search result is received from the search unit 13 (S1006). When the search result is received (S1006: YES), the search result is displayed on the search result part (D1405) of the search request screen (S1007), and it waits for the search button or the like to be pressed again (S1003, S1004).

FIG. 11 is a flowchart illustrating an example of the operation of the search unit 13.
When the file search server 1 receives a search request from the client 3, the search unit 13 is activated.
The search unit 13 first analyzes the search condition expression included in the search request and determines whether or not to perform metadata search, that is, whether or not it is necessary to search by referring to the system metadata 221 or the like of the index 2 file 22. (S1101). For example, if the search condition expression is “full text = site title = work report”, it is determined that the search needs to be performed with reference to the title 222 a of the standard metadata 222. For example, when the search condition formula is “full text = on-site”, it is determined that only the full text search is performed without performing the metadata search.
When it is determined that the metadata search is to be performed (S1101: YES), the search unit 13 performs a search using the index 2 file 22 (S1102). Specifically, a condition related to the system metadata 221 and the like is extracted from the search condition formula, and an index 2 record 220 that matches the condition is selected (hereinafter referred to as “metadata matching record”). For example, when the search condition formula is “full text = site title = work report”, the index 2 record 220 in which “work report” is set in the title 222 a of the standard metadata 222 is selected.

After performing the metadata search (S1102) or when determining that the metadata search is not performed (S1101: NO), the search unit 13 performs a full-text search, that is, refers to the keyword 214 of the index 1 file 21. It is then determined whether it is necessary to search (S1103). For example, if the search condition formula is “full text = site title = work report” or “full text = site”, it is determined that the search is necessary with reference to the keyword 214.
When it is determined that the full text search is performed (S1103: YES), the search unit 13 performs a full text search using the index 1 file 21 (S1104). Specifically, when the search unit 13 extracts a full-text search condition from the search condition formula and performs a metadata search (S1102), the file path 221c of the metadata matching record and the search request unit 31 are searched. Together with the user ID received from the index 1 search unit 14. When the metadata search is not performed, the search unit 13 selects all the index 2 records 220 of the index 2 file 22 as metadata matching records, and sets the file path 221c of the record as a full-text search condition. And the user ID received from the search request unit 31 is passed to the index 1 search unit 14.
As will be described later, the index 1 search unit 14 performs a search by referring to the received full-text search conditions and the like, and finally searches for an index 2 record 220 (hereinafter referred to as “keyword match record”) that should be the search result. The file path 221c is passed to the search unit 13.

When it is determined that the full text search is not performed (S1103: NO), the search unit 13 sets the metadata matching record as a transmission target to the search request unit 31. On the other hand, when the full text search is executed (S1104), the keyword The matching record is a transmission target to the search request unit 31. The search unit 13 transmits each data item of the index 2 record 220 to be transmitted to the search request unit 31 (S1105).
The search part 13 complete | finishes a process after transmission.

FIG. 12 is a flowchart showing an example of the operation of the index 1 search unit 14.
For example, the index 1 search unit 14 searches the index 1 record 210 corresponding to each metadata matching record (S1201). Specifically, the index 1 search unit 14 refers to the index 1 record 210 in which the same value is set in the file path 212 for all the file paths 221c of the metadata matching records received from the search unit 13, and the record The access authority 213 determines whether the user ID received from the search unit 13 has access authority. If it is determined that the user has access authority, it is determined whether the keyword 214 of the record satisfies the full-text search condition received from the search unit 13.
The index 1 search unit 14 passes the file path 221c determined to satisfy the full-text search condition to the search unit 13 (S1202), and ends the process.

  By the way, when using various existing full-text search programs as a program for realizing the index 1 search unit 14, the index 1 creation unit 11 also uses a program corresponding to the index 1 search unit 14. . In this case, generally, the search target file 43 related to the index 1 file 21 and the search target file 43 related to the index 2 file 22 are different, and the index 1 file 21 depends on the full-text search conditions regardless of the index 2 file 22. Will be searched. Then, for example, even if the operator of the client 3 sets a file that is frequently used as the search target file 43 related to the index 2 file 22, if the operator of the client 3 requests only a full-text search, the index 1 file 21 Will be searched, so files that are not normally used will be displayed as search results.

  It may be preferable to search in this way, but it may not be preferable. Therefore, when only full-text search is performed, only the search target file 43 in which the index 2 file 22 is created is set as the search target (that is, only the file for which metadata has been created is set as the search target), or the index 2 Whether to search all the search target files 43 of the index 1 file 21 regardless of the file 22 (that is, whether to search for files for which metadata is not created) may be specified on the search request screen. .

  In this case, when it is specified that a file for which metadata has not been created is also set as a search target, the index 1 search unit 14 searches the index 1 file 21 according to the full-text search condition regardless of the index 2 file 22. To do. On the other hand, when it is specified that only the metadata created file is to be searched, the search unit 13 also searches for the index 2 file even when the metadata search is not requested (NO in S1101 of FIG. 11). All 22 index 2 records 220 are selected as metadata matching records, and the file path 221c of the record is passed to the index 1 search unit 14 together with the full text search condition and the user ID received from the search request unit 31.

<Supplementary explanation about the virtual classification dynamic generation function>
FIG. 22 is a flowchart illustrating an example of the operation of the search unit 13 and the like in the virtual classification dynamic generation.
First, the search unit 13 analyzes a search condition expression included in the search request and determines whether or not the metadata search is performed (S1301).
When it is determined that the metadata search is performed (S1301: YES), the search unit 13 searches the keyword dictionary file 25 (S1302). Specifically, the keyword dictionary record 250 in which the keyword for metadata search is set in the keyword 253 of the keyword dictionary file 25 is selected (hereinafter referred to as “keyword dictionary matching record”). For example, when the search condition expression is “title = design document”, the keyword dictionary record 250 in which “design document” is set in the keyword 253 of the keyword dictionary file 25 is selected.
When the metadata search keyword exists in the keyword dictionary file 25 (S1303: YES), the search unit 13 searches the keyword related definition file 26 with the keyword ID 251 of the keyword dictionary matching record (S1304). Specifically, when the keyword ID 251 of the keyword dictionary match record is “1”, the keyword related definition record 260 in which “1” is set in the keyword ID 261 of the keyword related definition file 26 is selected (hereinafter referred to as “keyword related match”). Record ").

When the keyword ID 251 of the keyword dictionary matching record exists in the keyword related definition file 26 (S1305: YES), the search unit 13 sets the related keyword ID 262 by the number of keyword IDs set in the related keyword ID 262 of the keyword related matching record. The keyword dictionary file 25 is searched with the keyword ID set to (S1306). Specifically, when “2, 3, 4” is set in the related keyword ID 262 of the keyword related matching record, “2”, “3”, or “4” is set in the keyword ID 251 of the keyword dictionary file 25. The selected keyword dictionary record 250 is selected (hereinafter referred to as “related keyword dictionary matching record”).
Further, although not shown, when the value set in the maximum keyword length 263 of the keyword related match record is “1” or more, the related keyword dictionary match record is the keyword whose maximum keyword length is in the related keyword dictionary match record. Narrow down to records that are less than or equal to the length 263 value.
Further, although not shown, when the language condition 264 of the keyword related match record is set to “0 (only the same language)”, the related keyword dictionary match record is set to the language code 252 of the related keyword dictionary match record and the keyword dictionary. Only records with the same language code 252 of matching records are narrowed down.

The search unit 13 creates a virtual classification definition record from the metadata search keyword and the keyword dictionary record 250 selected and narrowed down by the above-described method (S1307).
Specifically, the search unit 13 sets the metadata search keyword to the display name 232 of the virtual classification definition record 230 and creates the highest virtual classification definition record 230. For example, when the metadata search condition is “title = design document”, the virtual classification ID 231, the display name 232, the condition 233, and the upper virtual classification ID 234 are “1”, “design document”, “no condition”, “ 0 (no higher rank) ”is set to create a virtual classification definition record 230 (hereinafter referred to as“ dynamically generated route virtual classification record ”).
Next, the search unit 13 creates a virtual classification definition record 230 subordinate to the dynamically generated root virtual classification record from the related keyword dictionary match record. For example, when the metadata search condition is “title = design document” and the keywords related to the metadata search keyword “design document” are “specification document” and “definition document”, the virtual classification ID 231, display name 232, condition 233 and upper virtual classification ID 234 are set to “2”, “specification”, “title = specification”, “1”, respectively, and “3”, “definition”, “title” respectively. = Definition document ”,“ 1 ”is set, and a virtual classification definition record 230 is created (hereinafter referred to as“ dynamically generated virtual classification record ”). At this time, the condition 233 creates and sets a condition “metadata search condition item (“ title ”) = related keyword (“ specification ”or“ definition ”)”. Further, the value “1” of the virtual classification ID 231 of the dynamically generated root virtual classification record is set as the upper virtual classification ID 234.

The search unit 13 creates a virtual classification definition based on the dynamic generation root virtual classification record and the dynamic generation virtual classification record. When the creation of the virtual classification definition is completed, the search unit 13 transmits the created virtual classification definition to the search request unit 31 (S1308). Note that the search unit 13 may store only the virtual classification definition file including the dynamic generation root virtual classification record and the dynamic generation virtual classification record in the memory of the file search server 1 without storing the virtual classification definition file in the storage device 2.
The search request unit 31 displays the received virtual classification definition as a tree on the virtual classification display unit (D1402) of FIG. 23, as with the virtual classification function.

<Supplementary explanation about processing when performing both metadata search and full-text search>
In the file search system of the first embodiment, as described above, the search is performed using the index 1 file 21 only when the operator of the client 3 requests a full text search. By the way, since the processing time required for full-text search is generally longer than when full-text search is not performed, the waiting time until the search result is displayed after the operator of the client 3 makes a search request is increased. Therefore, it is desirable to prevent the operator of the client 3 from waiting for a long time unexpectedly.
Hereinafter, when both the metadata search and the full-text search are performed (hereinafter referred to as “composite search”), the operation of the search unit 13 and the like when performing the above countermeasure will be described.

FIG. 16 is a flowchart illustrating an example of the operation of the search unit 13 and the like in the complex search.
S1650 to S1655 in FIG. 16 show details of processing performed by the search unit 13 in S1104 to S1105 in FIG. 11 during complex search. In S1601 to S1607, the search request unit 31 corresponds to the processing shown in FIG. The details of the processing performed in S1006 to S1007 are shown.

The search unit 13 compares the number of search results, that is, the number of records of metadata matching records searched in the metadata search, with a predetermined number (hereinafter referred to as “maximum search number”) (S1650).
When the number of metadata matching records, that is, the number of full-text search targets is larger than the maximum number of searches (S1650: YES), a message for confirming whether or not processing is continued is transmitted to the search request unit 31 (S1651). It waits until receiving a confirmation result as to whether or not to continue the search process from the search request unit 31 (S1652).

When receiving a message confirming whether or not to continue the search process from the search unit 13, the search request unit 31 displays the message on the display device 33 and determines whether or not to continue the search process to the operator of the client 3. Is requested (S1601). Specifically, for example, a confirmation message and buttons “Continue Search” and “Cancel” may be displayed on a pop-up screen, and wait until any button is clicked.
When the operator of the client 3 gives an instruction to cancel the search by clicking the “Cancel” button or the like (S1602: NO), the search request unit 31 transmits an instruction “Cancel Search” to the search unit 13 to perform the process. The process ends (S1603). Accordingly, the search request unit 31 waits for the search button or the like to be pressed again on the search request screen without displaying the search result (S1003 and S1004 in FIG. 10).
When the operator of the client 3 instructs to continue the search by clicking the “Continue Search” button or the like (YES in S1602), the search request unit 31 transmits a “Continue Search” instruction to the search unit 13, Although not clearly shown in the figure, it waits until a search result is received from the search unit 13.

When the search unit 13 receives a “search stop” instruction or a “continue search” instruction from the search request unit 31, the search unit 13 changes processing according to the received instruction (S1652). That is, when the “search stop” instruction is received (S1652: NO), the process is terminated. On the other hand, when the “Continue Search” instruction is received (S1652: YES), the file path 221c of the record corresponding to the maximum number of searches among the metadata matching records, the full-text search condition, and the user ID received from the search unit 31 are displayed. The full text search is executed by passing it to the index 1 search unit 14 (S1653).
If the number of full-text search targets is equal to or less than the maximum number of searches (S1650: NO), the file path 221c of the metadata matching record, the full-text is sent without sending a message to confirm whether or not processing is continued to the search request unit 31. The search condition and the user ID received from the search request unit 31 are transmitted to the index 1 search unit 14 to execute a full text search (S1653).

When the full-text search is completed, the search unit 13 transmits each data item of the keyword match record to the search request unit 31 (S1654). At this time, the identification of whether or not the full-text search has been executed for all the metadata matching records or the metadata matching records for which the full-text search has not been executed remains.
Next, the search unit 13 determines whether additional display is possible in the search result unit (D1405) (S1655). More specifically, a record that has not been subjected to full-text search remains in the metadata matching record, and the cumulative value of the search results transmitted to the search request unit 31 is displayed in the search result portion (D1405) of the search request screen (D1401). When the number of cases that can be displayed is smaller than a predetermined number (hereinafter referred to as “maximum number of cases to be displayed”) (S1655: YES), a confirmation result as to whether or not to continue the search process is received from the search request unit 31 again. (S1652). On the other hand, when additional display is not possible in the search result part (D1405) (S1655: NO), the search part 13 ends the process.

  When receiving the search result from the search unit 13, the search request unit 31 displays the search result in the search result unit (D1405) (S1605). As described above, as long as additional display is possible in the search result part (D1405), full-text search for metadata matching records is repeatedly executed. Therefore, the search result is additionally displayed in the search result part (D1405). For example, when the result of the first full-text search is 3 and the result of the second full-text search is 4, the search result portion (D1405) displays 7 search results.

  Next, the search request unit 31 still has a metadata matching record that has not been executed yet. (As described above, the search unit 13 has executed a full text search for all metadata matching records, or a full text search. It is determined whether or not additional metadata can be additionally displayed in the search result part (D1405). If there is a metadata matching record that has not been executed in full text search and additional display is possible in the search result part (D1405) (S1606: YES), the search request part 31 sends a message for confirming whether or not to continue processing. The information is displayed on the display device 33 (S1607), and the operator of the client 3 is requested again for a response as to whether or not to continue the search process (S1602).

  On the other hand, if the search request unit 31 has performed full text search for all metadata matching records or cannot additionally display the search result unit (D1405) (S1606: NO), the process ends. Accordingly, the search request unit 31 displays the search results so far in the search result unit (D1405) and waits for the search button or the like to be pressed again (S1003 and S1004 in FIG. 10).

As described above, when the full-text search target number is larger than the maximum search number, the operator of the client 3 is inquired as to whether or not to continue the search process, and when “continue search” is instructed, the full-text search is performed. Therefore, if the search time is likely to be long, the operator of the client 3 can stop the search process once and perform a search by further narrowing down the metadata search conditions, for example.
Further, the full-text search is repeatedly performed in units of the maximum number of searches, and each time a full-text search is performed, the search results are additionally displayed. Therefore, the operator of the client 3 can retrieve the search results one after another in a relatively short time. Can be confirmed.

<Supplementary explanation about related search>
FIG. 17 illustrates the display contents of the search result part (D1405) included in the search request screen (D1401).
In FIG. 17, the file name and file path are displayed. In addition to this, data items such as the index 2 record 220, such as the title and the document description date, may be displayed, for example, the index 2 record 220. Alternatively, a part of the contents of the search target file 43 may be set and displayed.
In addition, the name set in the display name 242 of the related definition record 240 is displayed in the related search instruction section (D1701) surrounded by a dotted line in FIG. In the example of FIG. 17, there are two related definition records 240 in which “title” and “description date” are set in the display name 242, respectively, and this display name 242 is displayed.

In this state, when the operator of the client 3 clicks, for example, a part displayed as “title”, the search request unit 31 requests the search unit 13 to perform a related search related to “title”. Specifically, although not displayed on the display device 33, the file ID 221a related to the search result file and the related definition ID 241 related to the “title” received from the search unit 13 together with the search result are searched together with the related search request. To the unit 13.
Upon receiving the related search request, the search unit 13 refers to the condition 243 of the related definition record 240 in which the received related definition ID 241 is set, searches the index 2 record 220 according to the condition set in the condition 243, and searches The result is transmitted to the search request unit 31.

For example, there is a related definition record 240 in which “1”, “title”, and “title 222a are equal to the search result” are set in the related definition ID 241, display name 242, and condition 243, respectively. There are four index 2 records 220 including the report “, and each file name 221b is“ work report 1.doc ”,“ work report 2.doc ”,“ work report 3.doc ”and“ report ”. Suppose that “work report 1.doc” is displayed on the display device 33 as a search result. In this case, the text “title” is displayed in the related search instruction section (D1701). When the operator of the client 3 clicks on the “title” character, the search request unit 31 transmits the file ID and the related definition ID (“1”) of “work report 1.doc” to the search unit 13.
Then, the search unit 13 refers to the condition 243 of the related definition record 240 whose related definition ID 241 is “1”, and “Title 222a is equal to the search result” is set. The title 222a of the “work report 1.doc” is acquired, and the work report 2.doc “,” “work report 3. Three files “doc” and “report.doc” are searched, and a search result (for example, a file name and a file path) is transmitted to the search request unit 31. Then, the search request unit 31 displays the related search result in the search result unit (D1405) as shown in FIG.
Note that the number of related search results displayed may not be particularly limited. For example, only a maximum of 5 items may be displayed, and when the number of related search results exceeds 5, the number may be displayed on a separate screen.

<Other Embodiments of File Search System>
The file search system according to the present invention is not limited to the first embodiment, and can be various embodiments. One example will be described below.

FIG. 19 shows an example of the configuration of a file search system according to the second embodiment of the present invention.
In the second embodiment, unlike the first embodiment, the file search server 1 does not include the index 1 creation unit 11 and the index 1 search unit 14.
Instead, the file search server 5 is communicably connected to the client 3, the file server 4, the WEB server 7, and the file search server 1 through the communication line 9. The file search server 5 is a computer and includes a CPU, a main storage device, and a storage device 6 (not shown). A predetermined program is stored in the main storage device of the file search server 5. When the CPU executes the instruction code, the functions of the index 1 creation unit 11 and the index 1 search unit 14 are realized. The index 1 file 21 is not stored in the storage device 2 of the file search server 1 but is stored in the storage device 6 of the file search server 5.
The major differences in configuration between the second embodiment and the first embodiment are as described above.

  In the file search system of the second embodiment, a client 3, a file server 4, a WEB server 7, a file search server 1, and a file search server 5 are connected to each other via a communication line 9 so that they can communicate with each other. With such a configuration, for example, when a certain organization stores a file in the file server 4 installed in the data center, the index 2 file 22, the virtual classification definition for the search target file 43 stored in the file server 4 is stored. By creating the file 23, the related definition file 24, the keyword dictionary file 25, and the keyword related definition file 26, metadata search, virtual classification display, physical folder display, and related search are possible, but the organization does not manage them. For the search target file 72 stored in the storage device 71 of the WEB server 7, the index 1 creation unit 11 of the file search server 5 performs, for example, a so-called WEB crawling via a WEB server program such as Apache (Apache is a registered trademark). Index Create a scan 1 file 21, it is possible to enable a full-text search.

  In the second embodiment, the file search server 5 does not need to be set up by an organization that performs file search, and performs a full-text search or the like by using a file search server function that exists in the world. Therefore, it is possible to construct a search system with high flexibility and expandability.

As described above, according to the present invention, it is possible to efficiently search for a file in a file search environment in which there are files that should be capable of full text search and files that do not require full text search.
For example, for a file such as a document created in an organization, an index 2 file 22 may be created and a metadata search may be performed. By searching the index 2 file 22 using the metadata search keyword, the number of files displayed as a search result can be suppressed. Even if the file is stored in the wrong folder, the file can often be searched if there is a small-capacity index 2 file 22 storing the purpose of use of the file, the title, and the like.
On the other hand, files obtained from outside the organization, such as patent documents and technical papers, and files existing on a server outside the organization, such as a WEB server, may be referred to for purposes that were not intended at the time of acquisition. In many cases, full text search using keywords is required. For such a file, an index 1 file 21 may be created and a full text search may be performed.

DESCRIPTION OF SYMBOLS 1 ... File search server, 11 ... Index 1 creation part, 12 ... Index 2 creation part, 13 ... Search part, 14 ... Index 1 search part, 2 ... Storage device, 21 ... Index 1 file, 22 ... Index 2 file, 23 ... virtual classification definition file, 24 ... related definition file, 25 ... keyword dictionary file, 26 ... keyword related definition file, 3 ... client, 31 ... search request section, 32 ... input device, 33 ... display device, 4 ... file server, DESCRIPTION OF SYMBOLS 41 ... File management part, 42 ... Storage device, 43 ... Search object file, 5 ... File search server, 6 ... Storage device, 7 ... WEB server, 71 ... Storage device, 72 ... Search object file, 9 ... Communication line

Claims (5)

A file acquisition means for acquiring a first search target file group consisting of a plurality of search target files and a second search target file group consisting of a plurality of search target files from the storage device;
For each search target file included in the first search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored Index 1 file storage for creating an index 1 record having a specified file path and a plurality of keywords extracted from the contents of each search target file at a predetermined timing, and storing the index 1 file including each index 1 record Means,
For each search target file included in the second search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored are shown. Index 2 file storage for creating an index 2 record having a file path and one or more metadata representing the contents of each search target file at a predetermined timing and storing an index 2 file including each index 2 record Means,
Keyword search means for searching for each index 1 record included in the index 1 file and acquiring an index 1 record satisfying the search condition;
Metadata search means for searching for each index 2 record included in the index 2 file and obtaining an index 2 record satisfying the search condition;
A file search apparatus comprising: A virtual classification display name indicating the name of the virtual classification, an index 2 record presence / absence information indicating whether any index 2 record included in the index 2 file may be classified into the virtual classification, and the index 2 file In order to identify the classification condition for classifying the search target file corresponding to each index 2 record included in the virtual classification into the virtual classification, the upper hierarchy presence / absence information indicating the presence / absence of the upper hierarchy virtual classification, and the upper hierarchy virtual classification Virtual classification definition file storage means for storing a virtual classification definition file including a plurality of virtual classification definition records having higher layer identification information;
Based on the index 2 record presence / absence information, the classification condition, the upper hierarchy presence / absence information, and the upper hierarchy identification information of each virtual classification definition record included in the virtual classification definition file stored by the virtual classification definition file storage unit, Virtual classification definition means for creating virtual classification definition information indicating a tree-like hierarchical structure constituted by a classification and a search target file corresponding to each index 2 record;
Outputs the virtual classification definition information created by the virtual classification definition means, the virtual classification display name of each virtual classification constituting the tree-like hierarchical structure indicated by the virtual classification definition information, and the file name of each search target file Virtual classification output means,
The file search apparatus according to claim 1, further comprising: Keyword dictionary file storage means for storing a keyword dictionary file including a plurality of keyword dictionary records having a keyword and keyword identification information for identifying the keyword;
A plurality of keyword related definition records having keyword identification information of a keyword dictionary record included in the keyword dictionary file and keyword identification information of a keyword dictionary record included in one or more of the keyword dictionary files related to the keyword identification information. Keyword-related definition file storage means for storing a keyword-related definition file including:
When searching by the metadata search means, keywords included in the search conditions of the search, keywords and keyword identification information of each keyword dictionary record included in the keyword dictionary file, and included in the keyword related definition file Virtual classification dynamic generation means for generating the virtual classification definition record based on the keyword identification information of each keyword-related definition record and one or more keyword identification information related to the keyword identification information;
The file search apparatus according to claim 3, further comprising: The index 2 record includes index 2 record identification information for identifying the index 2 record,
One or more related definition records having related definition record identification information for identifying the related definition record, a display name indicating the name of the related definition, and a search condition acquisition condition for acquiring the search condition from the index 2 record. An associated definition file storage means for storing an associated definition file including:
Relation definition information output means for outputting relation definition record identification information and display name of each relation definition record included in the relation definition file storage means;
Receiving means for receiving related definition record identification information selected from each related definition record identification information output by the related definition information output means, and index 2 record identification information;
Search condition acquisition means for acquiring a search condition from an index 2 record including the index 2 record identification information received by the reception means based on a search condition acquisition condition of each related definition record included in the related definition file storage means; ,
Each index 2 record included in the index 2 file is searched, an index 2 record satisfying the search condition acquired by the search condition acquisition unit is acquired, and the file name and file path included in the acquired index 2 record Related search means for outputting
The file search apparatus according to any one of claims 1 to 3, further comprising: Computer
A file acquisition means for acquiring a first search target file group consisting of a plurality of search target files and a second search target file group consisting of a plurality of search target files from the storage device;
For each search target file included in the first search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored Index 1 file storage for creating an index 1 record having a specified file path and a plurality of keywords extracted from the contents of each search target file at a predetermined timing, and storing the index 1 file including each index 1 record Means,
For each search target file included in the second search target file group acquired by the file acquisition means, the file name that is the name of each search target file and the storage location in which each search target file is stored are shown. Index 2 file storage for creating an index 2 record having a file path and one or more metadata representing the contents of each search target file at a predetermined timing and storing an index 2 file including each index 2 record Means,
Keyword search means for searching for each index 1 record included in the index 1 file and acquiring an index 1 record satisfying the search condition;
Metadata search means for searching for each index 2 record included in the index 2 file and obtaining an index 2 record satisfying the search condition;
File search program to make it function.

Images