File system

The present invention relates to a file system of a computer device.

The capacity of data recording devices for computers such as magnetic disk devices is increasing year by year with the progress of technology. In recent years, there have been an increasing number of cases where long-term storage of data is obligatory due to laws and regulations. As a result, the amount of data handled by computer devices is increasing rapidly. In order to easily obtain the necessary files from such a large amount of data, a high-speed and high-performance file search technology is required.

As such a file search technique, an information retrieval system has been proposed which is provided with a dictionary of search keywords in addition to the document database and updates the dictionary each time a search is executed (Patent Document 1). When the user inputs a keyword, this information retrieval system displays another keyword that has been used in the same search formula as the keyword in the past to assist the user in inputting the keyword.

In addition, keywords are extracted and registered in advance from the document file, a search is performed for these keywords, and if this search fails, a search is performed for the entire text, and further, the keywords are registered according to the search results. A document retrieval device for updating is proposed (Patent Document 2). According to this document search device, even if the input keyword is not registered, a high-speed search using the registered keyword can be performed from the next time.<patcit num="1"><text>Japanese Patent Application Laid-Open No. 6-314296</text></patcit><patcit num="2"><text>Japanese Unexamined Patent Publication No. 9-198395</text></patcit>

<p> In the above search system and the like, since the document file and the information of the keyword related to the document file are recorded separately, even if the document file is moved to another computer system, the information of the keyword related to the document file does not move. Therefore, when the document file is moved to another computer system, the information of the keyword related to it cannot be used on the destination computer system.</p><p> Also, among the keywords related to one file, the keywords actually used to obtain the file are more important than the other keywords. However, in the above search system and the like, such a difference in the importance of keywords is not used for the search.</p>

<p> The present invention provides an attribute information area in which a registered keyword is stored as metadata of the data added to the file in a file system that manages a file composed of a file name and data recorded in the recording device. A file access control unit that controls access to a file and a file search unit that searches the file based on the input keyword are provided, and the file search unit uses the registered keyword and the input keyword. The file access control unit includes an attribute information search unit for comparing and searching, and a full-text search unit for comparing and searching the entire file with the input keyword, and the file access control unit is the result of the search by the full-text search unit. It is characterized by including an attribute information changing unit that changes the registered keyword based on the above.</p>

<p> According to the present invention, even if a file is moved to another computer system, the same high-speed keyword search as before the move can be performed.</p><p> In addition, the importance of the keyword can be reflected in the search result to improve the convenience of the user. Furthermore, the importance of keywords can be set and changed automatically.</p>

FIG. 1 is a block diagram of a computer system to which the file system of the present invention is applied.

The file server 103 is communicably connected to the client computer 101 via the network 102.

The file server 103 includes a processor device 104 and a disk device 105, and writes and reads data in accordance with a request from the client computer 101.

The processor device 104 includes a CPU (not shown), a memory (not shown), an interface for communicating with the client computer 101 via the network 102 (not shown), and the like, and receives a data write / read request from the client computer 101. It processes and manages the data recorded in the disk device 105. The file system 106 and the operation history table 109 (FIG. 5) are recorded in the memory (not shown) of the processor device 104. The file system 106 includes a file access program 107 (FIG. 2) and a file search program 108 (FIG. 3). These programs are executed by the CPU (not shown) of the processor device 104.

The disk device 105 is, for example, a magnetic disk device, which may be one disk drive or a disk array including a plurality of disk drives.

A file 110 composed of metadata 111 and body data 112 is recorded on the disk device 105. Here, the text data 112 is data requested to be written by the client computer 101. Further, the metadata 111 is management information added to the text data 112 by the file system 106, and includes the keyword information of the present invention. This keyword information will be described in detail later with reference to FIG.

FIG. 2 is an explanatory diagram of the configuration of the file access program 107.

The file access program 107 includes a file open routine 201, a file close routine 202, a file read routine 203, a file write routine 204, a metadata read routine 205, and a metadata change routine 206.

The file open routine 201 and the file close routine 202 will be described in detail with reference to FIGS. 6 and 7, respectively.

The file read routine 203 receives a read request from the client computer 101 and reads the file 110 from the disk device 105. The file read routine 203 is also called from the full-text search routine (FIG. 10).

The file write routine 204 receives a write request from the client computer 101 and writes the file 110 to the disk device 105.

The metadata read routine 205 is called from the metadata search routine (FIG. 11) to read the contents of the metadata 111 of file 110.

The metadata change routine 206 is called from the full-text search routine (FIG. 10) and the file close routine 202 (FIG. 7) to rewrite the contents of the metadata 111 of the file 110.

FIG. 3 is an explanatory diagram of the configuration of the file search program 108.

The file search program 108 includes an advanced search routine 301, a simple search routine 302, a full-text search routine 303, and a metadata search routine 304. These will be described in detail in FIGS. 8, 9, 10 and 11, respectively.

FIG. 4 is an explanatory diagram of the contents of the metadata 111.

The metadata 111 includes the attribute information of the file name 403, the owner 404, the creation time 405, the keyword 406, and the important keyword 407. Here, the attribute name 401 is the name of each attribute, and the attribute value 402 is the value recorded as each attribute.

The file name 403 represents the name of the file 110 to which the metadata 111 is attached. In FIG. 4, the attribute value 402 of the file name 403 is "patent.doc". This means that the name of the file 110 to which this metadata 111 is attached is "Patent.doc".

Owner 404 represents the owner of file 110 to which its metadata 111 is attached. Here, the owner may be a person who has requested the writing of the file 110, a person who has access authority to the file 110, or a person who has created or changed the text data 112 of the file 110. Further, the attribute value 402 of the owner 404 may be any information that identifies the owner, and may be, for example, an ID when the owner accesses the network 102.

The creation time 405 represents the time when the file 110 to which the metadata 111 is attached is created.

Keyword 406 is a search keyword that hits file 110 (ie, obtained file 110 as a result of the search) and is registered by the metadata change routine 206 (Figure 10). In the example of FIG. 4, the keyword 406 includes "patent". This means that the file 110 was hit as a result of the user performing a search using "patent" as a keyword (that is, the file 110 was included in the file 110 obtained as a result of the search). means.

When the file 110 is newly created, the contents of the keyword 406 or the important keyword 407 of the other file 110 in the disk device 105 are referred to, and the contents equivalent to these are the contents of the newly created file 110. If it is included in the text data 112, the content may be registered in the keyword 406 of the newly created file 110.

Further, the history information of the file 110 (for example, the past name of the file 110 or the name of the directory in which the file 110 was stored in the past) may be registered in the keyword 406 of the file 110.

The important keyword 407 is a particularly important search keyword that hits the file 110, and is registered by the file closing routine 202 (Fig. 7).

Normally, the target file 110 is not always included in the file 110 obtained as a result of the keyword search, and even if the target file 110 is included, not all of the target files 110 are included. If the search result using a certain keyword includes the target file 110, the keyword is an important keyword for searching the target file 110. Such a keyword is registered as an important keyword 407.

In the example of Figure 4, the important keyword 407 includes "Web". This means that as a result of the user searching using "Web" as a keyword, the file 110 was hit and the file 110 was the target file.

It should be noted that the fact that the file 110 is the target file may be determined by the user browsing the file 110 for a predetermined time (for example, 10 minutes) or longer (FIG. 7). This is because it is highly possible that the file that the user who performed the search browsed for a long time after performing the search was the target file.

Further, among the files 110 hit by the search, the user himself / herself instructs the target file 110, and by inputting the instruction, the keyword 406 or the important keyword 407 may be registered only for the target file 110. For example, a list of files 110 hit by the search is displayed using a graphical user interface (GUI), and the user indicates the desired file 110 by checking the check box provided for each file 110. The instruction may be input by operating the "update" button provided on the screen (for example, clicking the mouse) (not shown).

Further, the user may manually register the keyword 406 and the important keyword 407.

Further, an access counter (not shown) may be provided for each of the keywords registered in the keyword 406 and the important keyword 407, and the importance of each keyword may be determined by the value of the access counter. For example, the access counter for each keyword is incremented by 1 each time the file is hit in a search using that keyword, and when comparing the access counter values for the keywords, the keyword with the larger access counter value is added. Judge as an important keyword.

Further, the keyword 406 and the important keyword 407 may be described in XML (eXtensible Markup Language) format.

FIG. 5 is an explanatory diagram of the operation history table 109.

The operation history table 109 is a table in which history information of operations performed by the user on the file 110 in the past is recorded, and whether or not the file 110 obtained as a result of the search is viewed by the user who performed the search. Or, it is used to obtain information about the time required for browsing and to judge the importance of the keyword. The operation history table 109 is referred to by the file open routine 201 and the file close routine 202, and is updated by the file open routine 201, the file close routine 202, and the metadata search routine 304.

Information is registered in the operation history table 109 for each entry. One entry consists of time 501 and its corresponding operations 502, user 503, keyword 504 and filename 505.

Time 501 is the time when the operation 502 is performed.

Operation 502 represents the type of operation performed. Search indicates that a search has been performed (Fig. 11), and open indicates that file 110 has been opened (Fig. 6).

User 503 is the name of the user who performed the operation 502.

The keyword 504 is a search keyword used in the search when the operation 502 is "search". Further, the keyword 504 is a search keyword used in the search obtained the file 110 when the operation 502 is "open".

The file name 505 is the name of the file 110 that is the target of the operation 502. However, when the operation 502 is "search", the target file 110 is not specified, so the file name 505 is left blank.

In the example of FIG. 5, the user "shimizu" (503) performs a search using "XML" as a keyword (504) at 12:10 (501) on February 2, 2002 (502). Then, at 12:15 (501) on February 2, 2002, the same user "shimizu" (503) opened "Specification.doc" (505) among the files obtained by the search. (502). Furthermore, at 12:20 (501) on February 2, 2002, the same user "shimizu" (503) opened "Drawing.doc" (505) among the files obtained by the search (5). 502).

The entry whose operation 502 is "search" may be deleted after a certain period of time has elapsed from the time 501 of the entry. At this time, among the files 110 obtained by the search, the keyword 504 of the entry may be deleted from the keyword 406 of the file 110 that has never been opened.

FIG. 6 is a flowchart of the file open routine 201.

The file open routine 201 is a subroutine that opens the file 110 and is called by the client computer 101.

When the file open routine 201 is called by the client computer 101, the file open routine 201 gets the username U and the file name F from the calling client computer 101 (601).

Next, open file 110 with file name 403 in disk device 105 (602).

Next, in the operation history table 109, the entry E in which the operation 502 is "search" and the user 503 is U is searched (603). The entry E whose operation 502 is "search" is registered by the metadata search routine 304 (Fig. 11).

It then determines if the entry E exists (604). If entry E does not exist, it is determined that the file 110 is not the result of the search (ie, the action to open the file 110 was done independently of the search) and is new to the operation history table. Exit File Open Routine 201 without adding an entry.

On the other hand, if the entry E exists, the file 110 is determined to have been obtained as a result of the user U searching using the keyword 504 of the entry E, and the value K of the keyword 504 of the entry E is acquired. Then (605), add a new entry to the operation history table 109 (606), and exit the file open routine 201. This new entry sets the time 501 to the current time, the operation 502 to "open", the user 503 to U, the keyword 504 to K, and the file name 505 to F.

FIG. 7 is a flowchart of the file closing routine 202.

The file closing routine 202 is a subroutine that closes the open file 110 and is called by the client computer 101.

When the file closing routine 202 is called by the client computer 101, the file closing routine 202 obtains the username U and the file name F from the calling client computer 101 (701).

Then close file 110 with file name 403 F (702).

Next, from the operation history table 109, the entry E in which the operation 502 is "open", the user 503 is U, and the file name 505 is F is searched, and the value T at time 501 and the value K of the keyword 504 are obtained from the entry E. Then delete the entry E from the operation history table 109 (703).

Next, the time from time T to the current time is calculated and compared with a predetermined time (704). In the example of FIG. 7, this predetermined time is 10 minutes.

If the time from time T to the current time is not longer than 10 minutes, it is determined that the file 110 was not the target file because the time spent browsing the file 110 is short, and the file without changing the metadata 111 is determined. Exit the close routine 202.

On the other hand, when the time from the time T to the current time is longer than 10 minutes, it is determined that the file 110 is the target file because the file 110 has been browsed for a long time. Since keyword K is an important keyword for searching the file 110, call the metadata change routine 206 and add the keyword K to the attribute value 402 of the important keyword 407 in the metadata 111 of the file 110 ( 705), the file closing routine 202 is terminated.

FIG. 8 is a flowchart of the detailed search routine 301.

The detailed search routine 301 is a subroutine that performs a detailed search by a keyword, and is called from the client computer 101. Specifically, the metadata search routine 304 is executed using the keyword, and then the full-text search routine 303 is executed regardless of whether or not the file 110 is obtained as a result.

When the advanced search routine 301 is called by the client computer 101, the advanced search routine 301 obtains the user name U, the search keyword K, and the keyword update prohibition flag X from the called client computer 101 (801).

Next, the metadata search routine 304 is called, and a search (metadata search) by the search keyword K is executed for the keyword 406 and the important keyword 407 of the metadata 111 (802).

Next, the full-text search routine 303 is called to execute a search using the search keyword K for the body data 112 (803).

The list of files 110 obtained as a result of executing the metadata search routine 304 and the full-text search routine 303 is then returned to the client computer 101 (804).

At this time, the file 110 including the search keyword K in the important keyword 407 may be listed at the top of the list. This is because such file 110 is likely to be the file for search purposes.

In this case, in the metadata search routine 304, the search targeting only the important keyword 407 is executed first, the list of the search results is created, and then the list of the search results targeting the keyword 406 and the full text are created. The list obtained by executing the search routine 303 may be added.

FIG. 9 is a flowchart of the simple search routine 302.

The simple search routine 302 is a subroutine that performs a simple search using keywords, and is called from the client computer 101. Specifically, the metadata search routine 304 is executed using the keyword, and then the full-text search routine 303 is executed only when the file 110 is not obtained as a result.

When the simple search routine 302 is called by the client computer 101, the simple search routine 302 obtains the user name U, the search keyword K, and the keyword update prohibition flag X from the called client computer 101 (901).

Next, the metadata search routine 304 is called, and a search (metadata search) by the search keyword K is executed for the keyword 406 and the important keyword 407 of the metadata 111 (902). Next, it is determined whether or not the file 110 is obtained as a result of this metadata search (902) (903). If one or more files 110 are obtained, the process proceeds to step 905 without executing the full-text search routine 303.

On the other hand, if no file 110 is obtained as a result of the metadata search (902), the full-text search routine 303 is called and a search using the search keyword K is executed for the body data 112 (904).

It then returns the list of files 110 obtained as a result of these searches to client computer 101 (905).

At this time, the file 110 including the search keyword K in the important keyword 407 may be listed at the top of the list. This is because such file 110 is likely to be the file for search purposes.

In this case, in the metadata search routine 304, the search targeting only the important keyword 407 is executed first, the list of the search results is created, and then the list of the search results targeting the keyword 406 and the full text are created. The list obtained by executing the search routine 303 may be added.

FIG. 10 is a flowchart of the full-text search routine 303.

The full-text search routine 303 is a subroutine that executes a keyword search on the body data 112 of the file 110 and returns the search results to the routine called as a list of files, and is called by the detailed search routine 301 or the simple search routine 302.

When the full-text search routine 303 is called by the detailed search routine 301 or the simple search routine 302, the full-text search routine 303 obtains the user name U, the search keyword K, and the keyword update prohibition flag X from the called routine (1001).

Here, the keyword update prohibition flag X is a flag provided to protect the privacy of the user, and when this value is "ON", the keyword 406 of the metadata 111 and the keyword 406 regardless of the search result. Do not update important keywords 407 and operation history table 109.

The value of the keyword update prohibition flag X may be input by the user by displaying a dialog box asking whether or not to update the keyword 406 or the like on the user's operation screen (not shown).

Next, for all the files 110 in the disk device 105, the file open routine 201 and the file read routine 203 are called to access the text data 112, and it is investigated whether or not the search keyword K is included in the text data 112. (1002).

Next, a list of files 110 hit by the search of the text data 112 (that is, files 110 containing the search keyword K in the text data 112) is created, and the list is returned to the routine that called the full-text search routine 303 (1003). ).

Next, it is determined whether or not the flag X is "ON" (1004), and if it is "ON", the full-text search routine 303 is terminated.

On the other hand, if the flag X is not "ON", the metadata change routine 206 is called for all the files 110 hit in the search of the body data 112, and the keyword 406 of the metadata 111 of each file 110 is called. Add the search keyword K to the attribute value 402 of (1005), and end the full-text search routine 303.

FIG. 11 is a flowchart of the metadata search routine 304.

The metadata search routine 304 is a subroutine that executes a keyword search on the metadata 111 of the file 110 and returns the search results to the routine called as a list of files, and is called by the advanced search routine 301 or the simple search routine 302. ..

When the metadata search routine 304 is called by the advanced search routine 301 or the simple search routine 302, the metadata search routine 304 obtains the user name U, the search keyword K, and the keyword prohibition flag X from the called routine (1101).

Next, for all the files 110 in the disk device 105, the metadata reading routine 205 is called to access the metadata 111, and the search keyword K is included in the attribute value 402 of the keyword 406 of the metadata 111 or the important keyword 407. Investigate whether or not (1102).

Next, a list of files 110 hit by the search for metadata 111 (that is, files 110 containing the search keyword K in the attribute value 402 of keyword 406 of metadata 111 or important keyword 407) is created, and the list is used as meta. Returns the data retrieval routine 304 to the calling routine (1103).

At this time, the file 110 including the search keyword K in the important keyword 407 may be listed at the top of the list. This is because such file 110 is likely to be the file for search purposes.

In this case, in the metadata search routine 304, the search targeting only the important keyword 407 is executed first, the list of the search results is created, and then the list of the search results targeting the keyword 406 is added. You may.

Next, it is determined whether or not the flag X is "ON" (1104), and if it is "ON", the metadata search routine 304 is terminated.

On the other hand, if the flag X is not "ON", a new entry is added to the operation history table 109 (1105) and the metadata search routine 304 is terminated. This new entry sets time 501 to the current time, operation 502 to "search", user 503 to U, and keyword 504 to K. However, if there is an entry in the operation history table 109 where the operation 502 is "search", the user 503 is U, and the keyword 504 is K, the time 501 of the entry is updated to the current time. ..

According to the present invention, since each file has metadata and the metadata includes keyword information for searching the file, high-speed keyword search for the metadata can be realized. Further, when the file is moved to another computer system, the keyword information is also moved together with the file, so that the keyword search can be performed at the same high speed after the move as before.

In addition, among the files obtained as search results, for the target file, the keyword used for the search is registered in the metadata as an important keyword, and the file obtained by searching for the important keyword is placed at the top of the search results. By displaying it, the convenience of the user is improved. Further, the registration of important keywords can be automatically performed based on the time when the user browses the file or the like.

Further, when the user selects not to update the keyword information or the like, information such as when and which file the user accessed does not remain, so that the privacy of the user can be protected.

In the above embodiment, an example in which the present invention is applied to a file system is shown, but the present invention can also be applied to a document management system, a storage device, and the like.

When the present invention is applied to a document management system, the file search program 108, the metadata reading routine 205, and the metadata change routine 206 may be implemented. In this case, the file 110 in FIG. 1 corresponds to the managed document containing the metadata 111 and the body data 112, the metadata read routine 205 reads the metadata 111, and the metadata change routine 206 modifies the metadata 111. , The file search program 108 performs a document search.

When the present invention is applied to a storage device, the above file system and operation history table may be implemented. In this case, the file server 103 of FIG. 1 is a storage device to which the present invention is applied.

The present invention can be used for an operating system or a database file system, and contributes to the realization of a high-speed and highly convenient search.

<figref num="1">It is a block diagram of the computer system to which the file system of this invention is applied.</figref><figref num="2">It is explanatory drawing of the structure of a file access program.</figref><figref num="3">It is explanatory drawing of the structure of a file search program.</figref><figref num="4">It is explanatory drawing of the content of metadata.</figref><figref num="5">It is explanatory drawing of the operation history table.</figref><figref num="6">It is a flowchart of a file open routine.</figref><figref num="7">It is a flowchart of a file closing routine.</figref><figref num="8">It is a flowchart of an advanced search routine.</figref><figref num="9">It is a flowchart of a simple search routine.</figref><figref num="10">It is a flowchart of a full-text search routine.</figref><figref num="11">It is a flowchart of a metadata search routine.</figref>

Code description

101 Client computer 102 Network 103 File server 104 Processor device 105 Disk device 106 File system 107 File access program 108 File search program 109 Operation history table 110 File 111 Metadata 112 Body data