JP2009093405A - System, method and computer program for data retrieval - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To make the designating order of search keys flexible, when the result of retrieval is drilled down, and to increase the number of search keys which can be designated at retrieval. <P>SOLUTION: A converter 104 reads a search-method definition file 160, which defines searching methods with respect to many data items, generates many search files 122, 122, ..., which are respectively correlated with the many data items, and stores these files. The file name of each of the search files 122 is a character string, in which a plurality of key items (or their identifiers) relating to each data item are connected. A search engine 102 retrieves a file name, which matches the search key designated by a user, from the population 123 of the search files 122, 122, .... When the result of the retrieval is to be gradually narrowed down, the search engine retrieves at each stage, the file names which match all the search keys, which are specified both at the present stage and at the previous stage from the filename population 123. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a system, a method, and a computer program for searching data at high speed.

  In a conventional general database, an index table in which a plurality of key items scheduled to be used as search keys at the time of database construction is stored as a file storing data items to be searched (hereinafter referred to as data items). Is created separately. The search for the target file using the search key is performed on the index table, and the search speed is high. However, when a search is performed using a key item that is not registered in the index table as a search key, all data item files are opened, all data items are read, and the data items are compared with the search key. The search speed is extremely slow because it is necessary. Therefore, in a conventional database, in order to perform a search at high speed, a search key must be selected within a range of key items registered in an index table created in advance.

  Also, if you want to narrow down the data items in the search results called drilldown (for example, first search the matching data items by specifying the search key “Kanagawa” in the category “prefecture” For example, specify the search key “Yokohama” for the category “city” and narrow down the data item group. By specifying multiple search keys with different categories in order, the data items in the search results are stepped. The order in which the search keys for each category can be specified is limited depending on the structure of the index table created at the time of database construction (for example, the search key for category "Province" is specified) Search key for category "City" can only be specified) In order to change the search key designation order, it is necessary to change the structure of the index table. Changing the structure of the index table means changing the table definition anew and rebuilding the database.

  In order to save the effort of creating such an index table, for example, Patent Document 1 discloses a code that systematizes the contents of each file to be searched (for example, a code that indicates a plurality of key items in a predetermined order). In other words, a method of searching for a target file by assigning each code as a file name to each file and searching for the file name using a search key is disclosed.

  Japanese Patent Laid-Open No. 2004-228561 gives a concatenation of a plurality of attributes of each image file to be searched in a predetermined order to each image file as a file name, and the directory name of the directory in which the image file is stored In addition, at least a part of the plurality of attributes is included, and at the time of searching, a search path name and a search file name are created based on a search key of a specific attribute specified by the user, and the search is performed. A method of searching for a target file by searching a directory name and a file name using a path name and a search file name is disclosed.

  As described above, according to the search method in which the key item is included in the file name or the directory name, the target file can be searched at high speed without depending on the index table.

JP 2003-6014 A JP 2006-106922 A

  However, Patent Documents 1 and 2 do not provide a clear solution regarding the problem of restriction of the search key designation order when the above-described search result is drilled down.

  In addition, since the length of the file name that can be handled by the OS of the computer is limited, the number of key items that can be incorporated into one file name is limited. As a result, the number of search keys that can be specified during a search is limited.

  Accordingly, an object of the present invention is to allow a search key designation order to be freely changed when a search result is dollar-downed.

  Another object is to increase the number of search keys that can be specified during a search.

According to one aspect of the present invention, a system for searching for a data item that matches a search key from among a number of data items,
A storage means for storing a character string group composed of a large number of character strings respectively associated with a large number of data items, wherein each character string in the character string group belongs to a plurality of categories related to each data item. Including the concatenation of a plurality of key items or the concatenation of identification codes of the plurality of key items;
A first search key selecting means for selecting a search key from the key items included in the character string population according to the user's selection;
Search means for issuing a search request for searching the character string group for a character string that matches the selected search key;
Key item display means for extracting a key item belonging to the category specified by the user from the searched character string obtained as a search result for the search request, and displaying the extracted key item;
A next search key is selected from the displayed key items according to the user's selection, and the selected next search key and the search key selected by the first search key selection means are selected. Second search key selection means for operating the search means again so as to issue a search request for searching for a character string matching both from the character string population;
The second search key selection means repeatedly operates a plurality of times in accordance with the selection of the user, and each time, the next search key selected this time, the search key selected in the previous time, and the first search key are selected. The search means is operated again so as to issue a search request for searching the character string group for a character string that matches all of the search keys selected by the search key selection means.

According to another aspect of the present invention, a method for searching for a data item that matches a search key from among a number of data items includes:
A storage step of storing a character string group composed of a number of character strings respectively associated with a number of data items, wherein each character string in the character string group belongs to a plurality of categories related to each data item Including the concatenation of a plurality of key items or the concatenation of identification codes of the plurality of key items;
A first search key selection step of selecting a search key from the key items included in the character string group according to a user's selection;
A search step for issuing a search request for searching the character string group for a character string that matches the selected search key;
A key list display step of extracting and displaying key items belonging to the category designated by the user from the searched character strings obtained as search results for the search request;
A next search key is selected from the displayed key items according to the user's selection, and the selected next search key and the search key selected by the first search key selection step are selected. A second search key selection step for causing the search means to operate again so as to issue a search request for searching for a character string matching both from the character string group;
A second search key selection step as means;
The second search key selection step is repeated a plurality of times in accordance with repetition of the user's selection, each time, the next search key selected this time, the search key selected in the previous time, and the first search key selection And a repetitive step of causing the search means to operate again so as to issue a search request for searching the character string group for a character string that matches all of the search keys selected in the step.

  A computer program according to still another aspect of the present invention causes a computer to execute the above-described search method.

  According to each of the system, method, and computer program for data search of the present invention described above, the search key designation order can be freely changed when the search result is pulled down.

  According to a preferred embodiment of the present invention, a number of search files associated with the number of data items are stored. Each of the large number of character strings is incorporated in the file name of each of the large number of saved search files. Then, each time the search key is selected, the next search key selected this time, the search key selected in the previous time, the first selected search key, and the unselected search keys of other categories A search file name formed by concatenating a wild card as a substitute for is created. Then, a search request for searching for the file name of the search file that matches the created search file name is issued.

  According to this configuration, since a desired data item can be searched using the file name search function provided by the OS of the computer, the search speed is high.

  According to the above-described embodiment, each key item or the identification code of each key item is further sandwiched from both sides by a predetermined delimiter in each character string incorporated in the file name of each search file. The identification codes of the plurality of key items or the plurality of key items are concatenated, and a predetermined prefix character and a suffix are respectively added to the beginning and the end of the plurality of key items or the identification codes of the plurality of key items concatenated. A character is added.

  According to this configuration, when creating a search file name by concatenating the selected search key and wildcard, you can always add a wildcard to the beginning and end of the search file name. When searching, it is possible to save the trouble of checking whether the search key in the search file name is at the head or tail, and the search speed is improved.

  Furthermore, according to the above embodiment, the reference information for each data item can be stored in each search file. Thus, each data item itself can be separated from the search file and can be present at any location inside or outside the computer.

  According to another preferred embodiment of the present invention, a key item list file is stored, and the plurality of character strings are stored separately from each other in the stored key item list file. According to this configuration, as described above, when a character string formed by concatenating the plurality of key items or their identification codes is incorporated into the file name of the search file, the file name length limit (This limits the number of key items that can be incorporated into a file name, that is, the number of search keys that can be specified), but there is no problem that the number of search keys that can be specified is limited by the limit of the file name length.

  In any of the above embodiments, when each of the character strings includes a concatenation of the identification codes of the plurality of key items related to the data items, the key items are used as the identification codes of the key items. It is possible to adopt a character string in which each character code constituting the character code is converted into a character string representing an M-ary number (M is a number greater than 2). Alternatively, as another method, an intermediate table file in which a large number of key items related to the large number of data items are registered is prepared, and each key item is identified on the intermediate table file as an identification code of each key item. You may employ | adopt the code | symbol for doing. In the latter method, for example, a code for specifying the position of each key item on the intermediate table file can be used as the identification code of each key item.

  By configuring the above character string using the identification code of the key item instead of the key item itself, it is highly possible that the number of characters representing each key item (search key) will be reduced. The number of search keys that can be increased.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows the overall configuration of a search system according to the first embodiment of the present invention.

  The search system 100 according to the present embodiment is typically (although not necessarily) implemented by a computer program and executed by one or more computers. As shown in FIG. 1, the search system 100 includes a search engine 102 and a converter 104.

  The search engine 102 has a function of searching for a specific data item from among a large number of data items using a file name search method. That is, as indicated by an arrow in FIG. 1, the search engine 102 inputs a search key designated by the user (process 130), and sends a search request for a file name that matches the designated search key to the OS of the computer OS. A list of search results (searched file names) of the file name search is received from the file system 110 (process 134) and sent to the user based on the received search result list. A display of predetermined information to be presented is created and displayed on the display device of the computer (process 136). Here, the predetermined information to be presented to the user displayed in the process 136 is, for example, a list of key items that the user can specify as the next search key.

  The converter 104 has a function of automatically generating in the storage device 120 a database 126 consisting of a group of files 122, 122,... That is, as shown by arrows in FIG. 1, the converter 104 stores search method definition information that defines a search method related to a large number of data items (for example, what key item in which category can be designated as a search key). Are automatically generated based on the input search method definition information, and a plurality of search files 122, 122,... Respectively associated with the plurality of data items are generated. The search files 122, 122,... Are stored in a predetermined directory of the storage apparatus 120 via the file system 110 (processing 140, 142). Prior to the file name search by the search engine 102 described above, as a preparatory work, the search file 122, 122,... Is automatically stored by the converter 104 (or manually input by the user). Must be stored within.

  The computer on which the search engine 102 is executed, the computer on which the converter 104 is executed, and the computer that manages the storage apparatus 120 may be the same machine or different machines.

  Each of the search files 122, 122,... Is associated with each of a plurality of data items that are final search targets for the user. In the present embodiment, as an example, each data item is the content itself of each file 122 or a part of its content. Therefore, by opening each file 122, the user can access each data item. Alternatively, as a modification, the reference information of each data item may be the content itself of each file 122 or a part of the content, instead of each data item itself. Here, the reference information of each data item is, for example, path information of each file storing each data item (hereinafter referred to as an entity file to be distinguished from the search file), a link to each entity file, or An alias for each entity file. In this case, each entity file (that is, each data item) may exist in a storage location (anywhere inside or outside the computer) different from each file 122. By opening each search file 122 and manipulating the reference information therein, the user can access each data item.

  The file name of each search file 122 is a character generated according to a certain rule based on a plurality of predetermined key items that can be used as search keys when searching for each data item associated with each file 122. A column, which is automatically determined by the converter 104. That is, in the search method definition information input to the converter 104 in the process 140 described above, a plurality of key items that can be used as the search key described above are defined as one item of the search method for each of the plurality of data items. Based on the definition, the converter 104 generates a file name of each search file 122 associated with each data item.

  The file format of each search file 122 may be fixed in advance (for example, fixed in text (.txt) format) or may vary depending on the contents of each file 122, but either However, this is also automatically determined by the converter 104. For example, in the search method definition information input in the processing 140 described above, the format of the search file is defined as one item of the search method for each data item, and the converter 104 determines each data item based on the definition. The file format of each search file 122 associated with the item is determined.

  Further, as described above, the contents of each search file 122 may be data including each data item itself or data including reference information of each data item. In any case, this is also automatically determined by the converter 104. For example, information for determining the contents of the search file for each data item (for example, each data item itself (for example, text data), or each data item is included in the search method definition information input in the processing 140 described above. (Reference information of the entity file that stores) is defined as one item of the search method, and based on the definition, the converter 104 determines the contents of each search file 122 associated with each data item (for example, The reference information of each data item or entity file defined in the search method definition information is used as the contents of the search file 122, or the contents of the entity file specified by the reference information are used as the search file 122. Etc.).

  The search engine 102 described above can search the file names of the search files 122, 122,... At high speed. That is, when the file system 110 of the OS generally stores the files 122, 122,... In the storage device 120 managed by the OS, the file names of the files 122, 122,. ,... Are entered into the file name management information 124, and the file name management information 124 is separated from the files 122, 122,. Stored in a special area used by the OS. The file name management information 124 is almost always cached on the main memory of the computer. When the file system 110 receives a file name search request from the search engine 102 in the process 132 described above, the file system 110 searches for the file name on the cached file name management information 124, and at that time, the search file 122. , 122,... Access to itself is not performed. Therefore, the file name search is performed at high speed.

  Taking advantage of the high speed of the file name search, the search engine 102 performs the search result obtained in the previous refinement stage (a refined file) when performing so-called drill-down refinement of search results. Instead of using the conventional general method of searching for the next refinement stage on the name group), instead of using the parent group of file names (that is, all search files 122, 122) at any refinement stage. ,... (Search for file names) is performed. In order to enable such a narrowing-down method, a wild card, particularly a wild card (typically “*”) that means an arbitrary character having an arbitrary length is effectively used. As a result, the restriction on the search key designation order for conventional drill-down is eliminated, and the search keys can be designated in any order.

  FIG. 2 shows a method for creating search files 122, 122,.

  As shown in FIG. 2, for example, one file 160 (hereinafter referred to as a search method definition file) 160 in which the search method definition information is registered is input to the converter 104. In the search method definition file 160, as described above, for each of a plurality of data items, a plurality of key items that can be used as search keys, a search file format, and a search file content are determined. Information (for example, each data item itself or reference information of an entity file storing each data item) is defined. In the specific example of the search method definition file 160 shown in FIG. 2, the first row includes category names of a plurality of key items that can be used as search keys (for example, “data item number”, “prefecture”, “city” ”,“ Machimura ”,“ Company ”,“ Last Name ”,“ First Name ”,“ Reference File Name ”, etc. Each row after the second row corresponds to each data item, There are sequentially defined character strings indicating key items belonging to a plurality of categories that can be used as search keys for each data item.

  In this specific example, the key item of the category “reference file name” for each data item is not only one of the key items that can be used as a search key, but also specifies the file format of the search file. (For example, the identifier “.txt” specifies the text format), or the entity file storing each data item may be specified (for example, the data item in the entity file “0001.txt”) (Data item number "001" is stored). As in this specific example, a specific key item on the search method definition file 160 may also serve to specify the file format or entity file of the search file, but as a modification, the file of the search file Information specifying the format and / or information specifying the entity file may be defined on the search method definition file 160 separately from the key item.

  It is for the user to create such a search method definition file 160, and to add, change and delete search keys and add, change and delete data items on such search method definition file 160. Easy. In short, database maintenance is easy. Therefore, it is also suitable for applications where data increases in minutes, such as a scheduler.

  The converter 104 receives the search method definition file 160, and automatically creates and stores a plurality of search files 122, 122,... Respectively associated with a plurality of data items according to the contents of the search method definition file 160. At that time, the converter 104 converts a plurality of key items of each data item input from the search method definition file 160 into a file name of the search file 122 associated with each data item according to a certain rule. A single search file may be generated for one data item, or a predetermined number of search files of two or more may be generated. When two or more search files are generated for one data item, the two or more search files are preferably placed in different folders (directories).

  There are many variations in the naming method of each search file 122. One of the simplest methods is the nomenclature referred to herein as “direct type”. When a single search file 122 is generated for each data item using a direct naming method, a character string that directly represents a plurality of key items for each data item input from the search method definition file 160 Are simply concatenated in a certain order (for example, in the order of arrangement on the search method definition file 160), and the file name of the single search file 122 is determined. In this case, the input character string may be used as it is in the file name, or when the input character string includes a 2-byte code character such as Japanese (a multi-byte character is a server or Even if you convert a 2-byte code character to a 1-byte code character with a natural language translation such as translation from Japanese to English, and use the converted 1-byte code character (because it may not be usable depending on the communication environment) Good. In any case, it is desirable to insert a predetermined delimiter (for example, an underscore “_”) between adjacent linked key items. By using the delimiter, the number of characters of each key item can be increased. Regardless, it can be determined to which category each of the key items incorporated in the file name belongs.

  Here, one specific example is shown. As an example, it is assumed that a plurality of key items for each data item are defined as follows on the search method definition file 160 in the CSV format. In the example below, the first line defines a category of key items, and each line below the second line defines a plurality of key items for each data item.

<Example of key item definition>
Number, Name, Date of birth, Date of birth, Date of birth, Gender, Implementation year, Implementation month, Implementation date, End month, Schedule year, Schedule month, Schedule date, Deadline, Reference file name
001, Wataru, Shoji, 1950, 04, 27, 2007, Male, 09, 10, 0910, 2007, 09, 10, 1020, 200709012330.txt
002, Mibu, Murata, 1967, 12, 06, Male, 2007, 09, 10, 0910, 2007, 09, 10, 1020, 20070801010.txt

  Based on this example of the key item definition, the file name of the search file associated with each data item is determined as follows, for example.

<Example of file name by direct naming method>
001_Wataru_Shoji_1950_04_27_m_2007_09_10_0910_2007_09_10_1020_200709012330.txt
002_Akio_Murata_1967_12_06_m_2007_09_10_0910_2007_09_10_1020_20070801010.txt

  In the above example of the direct type nomenclature, all of the plurality of key items of each data item are incorporated in the file name of the search file for each data item. In this case, search files for different data items may be stored in the same directory (folder).

  On the other hand, as a direct modification, search files for different data items may be stored in different directories (folders) using a hierarchical directory (folder). In this case, as in the above example, all of the plurality of key items of each data item may be incorporated in the file name of the search file of each data item. Alternatively, one of the plurality of key items may be included. A plurality of key items of each data item may be distributed and incorporated in the directory name and the file name, such that the part is incorporated in the directory name and the rest is incorporated in the file name. In short, in this alternative method, a path name combining a directory name and a file name is formed by concatenating a plurality of key items. Specific examples are as follows. In the example below, the slash “/” means a separator between directories (folders) at different levels.

<Examples of path names using another direct naming method>
Wataru_Shoji_1950_04_27_m / 2007/09 / 10_0910_10_1020_200709012330.txt
Akio_Murata_1967_12_06_m / 2007/09 / 10_0910_2007_09_10_1020_20070801010.txt

  The existence of a large number of search files in the same directory (folder) is not a practical problem for a computer. For example, a general personal computer marketed at present can perform a file name search without any trouble even if at least 100,000 search files exist in one folder. However, it is convenient for the user that the search files are stored by being classified into appropriately named hierarchical folders as in the latter method described above. For example, when this embodiment is used for a scheduler, for example, a plurality of folders for each person are prepared, a plurality of folders for each year are prepared in each person's folder, and a folder for each year is prepared. Prepare multiple folders for each month, save multiple search files in each month folder, and add the date + start time + end time + location information + reference file name to the file name of each search file Operation such as incorporation is possible.

  In this way, the method of classifying and storing search files in a hierarchical directory (folder) and incorporating a plurality of key items into the path name of each search file is the only direct naming method described above. Rather, it can be employed in other types of nomenclature described below. Thus, in the description of other types of naming below, the term “file name” may mean not only a literal file name but also a path name.

As another type of nomenclature for search files, there is a method called “encoding type” in this specification. In the encoding type nomenclature, each character of each input key item is 1 byte (automatically by the converter 104 without any natural language translation even if it contains multi-byte characters such as Japanese). Encoded to a character string. Then, a concatenation of a plurality of key items expressed by these encoded character strings is used as the file name of the search file. As an encoding method, a character code of each inputted character (for example, a character code based on UTF-8 (Unicode Translation Format-8)) is converted to an M number (M is a number larger than 2), for example, a hexadecimal number. 1-byte character group (0-9,
A method of converting to a character string using AF) is adopted. As a specific example, for example, character strings of a plurality of input key items are
Shoji, Wataru, 1950, 04, 27, if you are a man, according to the encoding type nomenclature, the file name is, for example,
5E8453F8_6E09_G31G39G35G30_G3H34_G32G37_7537.txt
It becomes. The encoding type nomenclature is convenient for applications that frequently search in Japanese characters, such as business card management and address book applications.

  As another type of file name naming method, there is a method called “composite type” in this specification. This naming method is a combination of the direct type and encoded type naming methods described above. That is, in the above encoding type method, when the input character is a single-byte alphanumeric character, there is a problem that it is encoded into a longer character string. In order to solve this problem, the composite type nomenclature classifies multiple key item categories into those that should be encoded and those that do not encode, the key items belonging to the former category are encoded, and after encoding. The key item belonging to the latter category is encoded, and the input character string is used as it is without generating the file name. In this case, the category in which the input character strings of all the key items belonging thereto are composed only of alphanumeric characters are not encoded, and the other categories are classified as encoded. Such a classification may be set in advance, for example, by adding a flag information indicating whether or not to encode for each category of the key item in the search method definition file 160, or a converter 104 may be automatically performed by analyzing the character type of the key item described in the search method definition file 160.

Specific examples of complex nomenclature include multiple defined key items
Shoji, Wataru, 1950, 04, 27, if it is a man, only the key items `` Shoji '', `` Wa '' and `` Men '' that are not expressed in half-width alphanumeric characters are encoded, so the file name is, for example,
5E8453F8_6E09_1950_04_27_7537.txt
It becomes. The compound nomenclature is convenient for applications that frequently search both Japanese characters and alphanumeric characters, such as business card management, address book, and scheduler.

  By the way, in the encoding type or composite type naming method described above, the character group that can be used for the notation of the key item after encoding is not limited to the hexadecimal character group described above. Use shorter hex character groups instead of hex character groups, eg 62 hex character groups (eg 0-9, az, AZ) to shorten file names or include them in file names It becomes possible to incorporate more key items. The search speed is further improved if the file name is shortened.

  As described above, in the database according to the present embodiment, a plurality of key items possessed by each data item are incorporated into a file name. As a result, the data item search and extraction mechanism is replaced with the file name search, which is the basic function of the OS. As a result, the search can be performed at high speed. In the database according to the present embodiment, it can be said that a file name is used as an index of the database. As a result, all formats that can be handled on a computer, such as images, moving images, word processing documents, and spreadsheets, are acceptable as the data format of the data item itself. Therefore, it is possible to handle data items in various data formats in an integrated manner. Further, the search file may be described in XML or HTML, so that not only the storage device in the computer but also external information can be comprehensively processed.

  Next, a search method performed by the search system 102 shown in FIG. 1 will be described.

  In this search method, a search file name is created using one or more search keys designated by the user, and a search file having a file name that matches the search file name is searched. Such a file name search is performed one or more times, and a file name list as a search result is obtained. Thereafter, the search file selected by the user from the file name list is opened, and the data item (final search purpose for the user) associated with the selected search file is obtained.

  As shown in FIG. 3, the search range in the above file name search is a mother group of file names of search files (file names of all search files) 123. Even when multiple search keys are specified in sequence and the file name group is narrowed down (drill-down), the search range at each stage is not the file name group narrowed down in the previous stage, but the file name mother. A group. The search file name used in the file name search is automatically created by the search engine 102 using one or more search keys designated by the user and the wild card “*”. That is, a search key for one or more categories designated by the user and a wildcard “*” for other categories are combined in the order corresponding to the arrangement order of the categories in the file name. The search file name is created.

  When drilling down the search results by specifying multiple search keys in sequence, the search key of the category specified by the user at this stage and all previous stages at that stage, and other The search file name is created by concatenating the wild card “*” for the category.

  A specific example of a method for creating a search file name by the search engine 102 is shown below. For example, suppose the file name population of the search file is as follows (the key item categories in each file name are “number”, “first name”, “last name”, “born” in order from the left) Year, birth month, birth date, gender, prefecture, and reference file name).

001_Wataru_Shoji_1950_04_27_m_Tokyo_0001.txt
002_Maiko_Shoji_1978_05_27_f_Tokyo_0002.txt
003_Satoru_Shoji_1952_01_20_m_Tokyo_0003.txt
004_Akio_Murata_1967_12_06_m_Tokyo_0004.txt
005_Hiroaki_Otsuka_1949_04_04_m_Tokyo_0005.txt
006_Yoshie_Kidokoro_1960_04_25_f_Tokyo_0006.txt
007_Issei_Shoji_1990_04_27_m_Kanagawa_0007.txt
008_Hanako_Kidokoro_1990_04_25_f_Tokyo_0008.txt

In this case, if the search key “04” in the category “Birth month” is specified, the search file name is
* _ * _ * _ * _ 04 _ * _ * _ * _ *. Txt
Thus, the designated search key “04” and the wild card as a substitute for the undesignated search key are concatenated.

If the search key “Shoji” for the category “Mr.” is specified, the search file name will be
* _ * _ Shoji _ * _ * _ * _ * _ * _ *. Txt
It becomes.

If both the above search keys are specified, the search file name will be
* _ * _ Shoji _ * _ 04 _ * _ * _ * _ *. Txt
It becomes.

  Also, when drilling down the search results, for example, the search key “04” for the category “born month” is specified in the first stage to narrow down the file name group, and the search key “ If you specify “Shoji” and further narrow down the file name group, and further refine with the search key “Tokyo” in the category “Prefecture” in the next stage, it will be as follows.

In the first stage, the search filename is
* _ * _ * _ * _ 04 _ * _ * _ * _ *. Txt
It becomes.

In the next stage, the search filename is
* _ * _ Shoji _ * _ 04 _ * _ * _ * _ *. Txt
Thus, both the search keys “04” and “Shoji” designated in this stage and the first stage are incorporated.

In the next stage, the search file name is
* _ * _ Shoji _ * _ 04 _ * _ * _ Tokyo _ *. Txt
Like this, all of the search keys “04”, “Shoji”, and “Tokyo” specified in this stage and the previous stage are incorporated.

  The search range is not a group of file names narrowed down in the previous stage, but a group of file names, not only in the first stage but also in any subsequent stage.

In the above specific example, one wildcard “*” is set for each search key of an unspecified category of the search file name, but this is not necessarily the case. As long as the category of the specified search key can be identified from the search file name, the search keys of a plurality of unspecified consecutive categories can be combined and one wild card “*” can be set. For example, the search file name exemplified in the first stage above,
* _ * _ * _ * _ 04 _ * _ * _ * _ *. Txt
For example,
* _04 _ * _ * _ * _ *. Txt
It can also be.

  As described above, in the file name search, the search range is a group of file names even when the search result is drilled down. There is no need to create an index table in advance. As a result, there are no particular restrictions on the categories and order of search keys that can be specified when drilling down on search results, and the user can freely determine them. Even if the search is performed from the mother group every time as described above, the search speed is high (unless the number of wild cards in the search file name is very large). In addition, the wild card can be used not only as a substitute for an unspecified search key as described above, but also for a general fuzzy search purpose (substitute for a fuzzy part in each search key).

  FIG. 3 shows an example of a GUI (Graphical User Interface) window provided by the search engine 102. 4 to 7 show a process flow of search performed by the search engine 102. Hereinafter, a search process performed by the search engine 102 will be described with reference to these drawings.

  As shown in FIG. 4, the search engine 102 prepares a search file name template in step 200. The search file name template is obtained by concatenating the same number of variables as the search key to be included in the search file name (the part indicated by the number in the figure) via the concatenation character “_”. Each variable can contain a search key or wildcard.

  In step 202, the search engine 102 retrieves all categories of key items read by the converter 104 from the first line of the search method definition file 160 (see FIG. 2) in the GUI window 170 as shown in FIG. Are displayed in a “key category list” box 172.

  In step 204, the user designates an arbitrary category from the “key category list” box 172. Then, the search engine 102 selects a category specified by the user, and displays the selected category in a “selected key category” box 174 in the GUI window 170. At the same time, the search engine 102 grasps the arrangement position in the file name of the key item to which the selected category belongs (for example, the number of the key item from the left of the file name) (for example, The search method definition file 160 is grasped on the basis of the result read by the converter 104), and the grasped arrangement position is entered in the variable n1.

  In step 206, the search engine 102 displays a “first key list” box 180 on the GUI window 170 and empties its contents. Here, the “first key list” box 180 is a box for displaying a list of key items that the user can first designate as a search key.

  In step 208, the search engine 102 requests the OS file system 110 to obtain a list of file name populations of search files. In steps 208 to 216, the search engine 102 sequentially checks all the file names in the file name parent group list.

  In the process of checking the file name population, in step 212, the key item at the position corresponding to the variable n1 in each file name is extracted, and the key item is put in the variable T. At 216, the contents of the variable T are added to the “first key list” box 180 (key items that have already been added are not added), and step 210 causes the above checking process to be included in the file name parent list. Repeat until the last file name.

  As a result, in step 218, all key items belonging to the category first designated by the user are extracted and displayed in a list in the “first key list” box 180. In other words, key items that the user can select as the first search key are displayed in a list in the “first key list” box 180.

  Reference is now made to FIG. In step 220, the user designates an arbitrary category different from the category previously selected from the “key category list” box 172 of the GUI window 170. Then, the search engine 102 selects a category specified by the user, and displays the selected category in a “selected key category” box 174 in the GUI window 170. At this time, the category selected this time is displayed below the first category already displayed in the “selected key category” box 174. At the same time, the search engine 102 grasps the arrangement position in the file name of the key item to which the selected category belongs (for example, the number of the key item from the left of the file name) (for example, The search method definition file 160 is grasped on the basis of the result read by the converter 104), and the grasped arrangement position is entered in the variable n2.

  In step 222, the user selects an arbitrary key item as a search key from within the “first key list” box 180 of the GUI window 170. Then, the search engine 102 puts the selected search key (key item) in the variable k1, and creates a search file name using the variable k1. That is, in the search file name template prepared in step 202, the variable at the placement position corresponding to the variable n1 is replaced with the variable k1, and the other variables are replaced with the wild card “*”. A file name is created.

  In step 224, the search engine 102 requests the OS file system 110 to search the file name population of the search file for a file name that matches the created search file name, and the file system 110 Returns the first file name list as the search result to the search engine 102. In the first file name list, file name groups narrowed down by a search key first designated by the user are described.

  In step 226, the search engine 102 displays a “second key list” box 182 on the GUI window 170 and emptyes its contents. Here, the “second key list” box 182 is a box for displaying a list of key items that can be designated as a search key second by the user.

  In steps 228 to 236, the search engine 102 sequentially checks all file names in the first file name list obtained as the first search result.

  In the process of checking the first file name list, in step 232, the key item at the position corresponding to the variable n2 in each file name is extracted, the key item is put in the variable T, and step 234 is executed. 236, the contents of the variable T are added to the "second key list" box 182 (the key item that has already been added is not added), and step 230 causes the above check processing to be performed in the first file name list. Repeat until the last file name.

  As a result, in step 238, all key items belonging to the second designated category included in the file name group narrowed down by the search key designated first by the user are displayed in the “second key list” box 182. Are listed. In other words, key items that the user can select as the second search key are displayed in a list in the “second key list” box 182.

  Reference is now made to FIG. After this, if the user does not specify the next category in step 240, the control proceeds to the flow shown in FIG. 7, but if the next category is specified, the control proceeds to step 242.

  In step 242, the user designates an arbitrary category different from the category selected so far from within the “key category list” box 172 of the GUI window 170. Then, the search engine 102 selects a category specified by the user, and displays the selected category in a “selected key category” box 174 in the GUI window 170. At this time, the category selected this time is displayed below the category selected so far, which is already displayed in the “selected key category” box 174. At the same time, the search engine 102 grasps the arrangement position in the file name of the key item to which the selected category belongs (for example, the number of the key item from the left of the file name) (for example, The retrieval method definition file 160 is grasped on the basis of the result read by the converter 104), and the grasped arrangement position is entered in the variable n3.

  In step 244, the user selects an arbitrary key item from the “second key list” box 182 of the GUI window 170 as a search key. Then, the search engine 102 puts the selected search key (key item) in the variable k2, and creates a search file name using the previous variable k1 and the current variable k2. That is, in the search file name template prepared in step 202, the variable at the placement position corresponding to the variable n1 is replaced with the variable k1, and the variable at the placement position corresponding to the variable n2 is replaced with the variable k2. By replacing the other variables with the wild card “*”, the search file name is created.

  In step 246, the search engine 102 requests the OS file system 110 to search the file name population of the search file for a file name that matches the created search file name, and the file system 110. Returns the second file name list as the search result to the search engine 102. In the second file name list, file name groups narrowed down by both the search key designated first by the user and the search key designated second are described.

  In step 248, the search engine 102 displays a “third key list” box 184 on the GUI window 170 and emptyes its contents. Here, the “third key list” box 184 is a box for displaying a list of key items that the user can designate as the third search key.

  In steps 250 to 258, the search engine 102 sequentially checks all the file names in the second file name list obtained as the second search result.

  In the process of checking the second file name list, in step 254, the key item at the position corresponding to the variable n3 in each file name is extracted, the key item is put in the variable T, and step 256 is performed. And 258, the contents of the variable T are added to the "third key list" box 184 (the key item already added is not added), and the above check processing is added to the second file name list in step 252. Repeat until the last file name.

  As a result, in step 260, all key items belonging to the third designated category included in the file name group narrowed down by both the first and second designated search keys by the user are displayed as “third A list is displayed in a “key list” box 184. In other words, key items that the user can select as the third search key are displayed in a list in the “third key list” box 184.

  Thereafter, as shown in step 262, the above-described processing from the category selection to the creation of the next “key list” box is repeated as many times as the number of refinements desired by the user. For example, in the example shown in FIG. 3, the narrowing down is repeated four times, thereby creating a “fourth key list” box 186.

  Reference is now made to FIG. It is assumed that the user does not desire further refinement at the stage where the refinement is performed up to the i-th time. At this time, the “i-th key list” box (“fourth key list” box 186 in the example of FIG. 3) is displayed in the GUI window 170. In this case, in step 264, the user selects an arbitrary key item as a search key from the “i-th key list” box. Then, the search engine 102 puts the selected search key (key item) in the variable ki, and creates a search file name using the variables k1 to ki so far. That is, in the search file name template prepared in step 202, the variables at the placement positions corresponding to the variables n1 to ni are replaced with the variables k1 to ki, and the other variables are replaced with the wild card “*”. By doing so, a search file name is created.

  In step 266, the search engine 102 requests the OS file system 110 to search the file name population of the search file for a file name that matches the created search file name, and the file system 110. Returns the i-th file name list as a search result to the search engine 102. In the i-th file name list, file name groups narrowed down by all the search keys designated from the beginning to the current time by the user are described.

  In step 268, the search engine 102 extracts key items of one or more predetermined categories from all file names in the i-th file name list, and searches the extracted key items for “search” on the GUI window 170. The “Result” box 188 is displayed. The key item extracted here is preferably in a category that makes it easy for the user to identify what the corresponding data item is. For example, if the data item is product data, the key items of the categories “product name” and “product number” may be suitable for it.

  In step 270, the user selects any key item from within the “search results” box 188. Then, the search engine 102 requests the OS file system 110 to open a search file corresponding to the selected key item, and obtains a data item associated with the search file. At this time, if the data item itself is stored in the search file, the data item can be obtained by opening the search file. On the other hand, if the reference information of the data item is stored in the search file, the search file is first opened to obtain the reference information of the data item, and then the reference information is used. The data item can be acquired.

  In step 272, the search engine 102 displays the acquired data item in a “data content” box 190 on the GUI window 170.

  According to the search process described above, search keys of various categories can be designated in an arbitrary order when drilling down a search result.

  In the embodiment described above, the number of key items constituting one file name is restricted by the limit of the length of the file name that can be handled by the OS. In addition, even if the length of the file name is less than or equal to the limit length, if the number of wildcards included in the search file name is large (about several tens or more), the file name search speed is high. There is a problem that gets worse. Furthermore, there may be a user who feels resistance that the number of search files corresponding to the number of data items is necessary. Further, when taking a backup of the database, it is necessary to copy as many search files as there are data items.

  The following describes some of the improved file name naming schemes that remedy these problems.

  First, a nomenclature called “table correspondence type” in this specification will be described. The table-based nomenclature provides an improvement over the problems of limiting the number of key items that can be included in a single file name, and degrading search speed due to heavy use of wildcards. In the table-corresponding naming method, an intermediate table for reducing the number of characters of each key item is interposed between the search method definition file 160 (see FIG. 2) and the search files 122, 122,. The intermediate table is created based on the search method definition file 160. In the intermediate table, key items defined by the search method definition file 160 and their categories are registered. In the intermediate table, how many key items exist for each category, and what category is registered at which position (which row position) in the table, and at which position (which column position) in each row It is configured so that it can be understood what key item is registered.

  This method can be further classified into two nomenclatures referred to herein as “column position type” and “matrix position type”, respectively.

  In the “column position type” nomenclature, when creating a file name, instead of concatenating multiple key items, the code indicating the column position in the intermediate table of each key item is assigned to the category of each key item. By concatenating in the order of the line positions, a file name is generated.

  Specific examples are shown below. For example, suppose that the CSV format search method definition file 160 defines key items as follows.

<Key item definition example>
Number, name, gender, age, year of birth, month of birth, date of birth, postal code, prefecture
01, Shoji, Wataru, Male, 57, 1950, 04, 27, 1620067, Tokyo
02, Shoji, Satoru, Male, 56, 1952, 01, 20, 1620067, Tokyo
03, Castle, Rye, Woman, 47, 1960, 04, 25, 1401234, Tokyo
04, Otsuka, Hiroaki, Male, 58, 1949, 04, 04, 1500001, Tokyo
05, Tanaka, Noriko, Woman, 40, 1966, 07, 13, 2220005, Aichi
06, Okamoto, Yukiko, Woman, 50, 1967, 09, 25, 3130088, Kanagawa Prefecture
07, Ito, Kumiko, Woman, 59, 1949, 10, 13, 2220000, Saitama

  From such a search method definition file 160, for example, the following intermediate table is created in the form of a CSV file.

<Example of intermediate table>
Number (7), 01,02,03,04,05,06,07
Mr. (6), Shoji, Castle, Otsuka, Tanaka, Okamoto, Itona (7), Wataru, Satoru, Yoshie, Hiroaki, Noriko, Yukiko, Kumiko Sex (2), Male, Female Age (7), 57, 56,47,58,40,50,59
Birth year (6), 1950, 1952, 1960, 1949, 1966, 1967
Birth month (5), 04,01,07,09,10
Birthday (5), 27,20,25,04,13
Postal code (6), 1620067,1401234,1500001,2220005,3130088,2220000
Prefectures (4), Tokyo, Aichi, Kanagawa, Saitama

  In this intermediate table, each key item registered there can be identified by the row position and the column position on this table. That is, the category to which each key item belongs is identified by the row position, and each key item belonging to each category can be identified by the column position in each row. Further, in this intermediate table, the number of key items (numbers in parentheses) for each category is known. In the above example (the “number” on the top line is a category that is not used as a search key, so this is ignored), the first line corresponds to the category “Mr.” and the key item of this category “Mr.” The number of is "01", "02", "03", "04", "05", "06" and "07", and the first character of each is (key item is Corresponding to columns 1, 4, 7, 10, 13, 16, 19 in the registered column range, the second row corresponds to the category “sex”, and the key of this category “sex” The number of items is 2, and they are “male” and “female”, and the first character of each item corresponds to the first column and the third column (within the column range where the key item is registered). You can see that This intermediate table file can be automatically created from the search method definition file 160 and saved by the converter 104 (see FIG. 2).

  The converter 104 uses the intermediate table as described above to create a search file group having the following file names from the search method definition file 160.

<Examples of file names for search files>
01_1_1_1_1_1_1_1_1_1.txt
02_1_2_1_2_2_2_2_1_1.txt
03_2_3_2_3_3_1_3_2_1.txt
04_3_4_1_4_4_1_4_3_1.txt
05_4_5_2_5_5_3_5_4_2.txt
06_5_6_2_6_6_4_3_5_3.txt
07_6_7_2_7_4_5_5_6_4.txt

  In each of the above file names, each code (for example, a number) sandwiched from both sides by a delimiter “_” indicates each key item. The position of each code in the file name (for example, what number from the left) is the row position in the intermediate table of the category to which each key item belongs (for example, what number category from the top) (in short, Category identification code), and the value of each code (for example, a number) is the column position in the intermediate table of the first character of each key item (for example, the number of the character from the left) (in short, each category The identification code of the key item in

In this case, the search file name is created by using a code indicating the column position in the intermediate table of the specified search key and a wild card as a substitute for the unspecified search key, and using the delimiter "_" In this way, the categories are linked in the row order of the categories in the intermediate table. In this method, it is not always necessary to set a wildcard one by one at each unspecified search key position, as long as the specified search key category can be identified. One wildcard can be set for a key block. For example, when the search key “male” of the category “gender” is specified, the search file name is
* _ * _ * _ 1 _ *. Txt
It can be.

  In this search file name example, the search key category is identified by the position from the left in the file name. Therefore, if all the search keys on the right side of the designated search key are not specified, one wild card can be set for all the search keys on the right side. In the case of this example, if the search key used more frequently is arranged on the left side, the number of wild cards to be set is reduced, so that the search speed is improved.

  Even when this “column position type” nomenclature is adopted, the search process described with reference to FIGS. 3 to 7 can be applied as it is.

  According to this “column position type” nomenclature, the file name can be shortened, or the number of key items that can be incorporated into a file name of a limited length can be increased. Further, as exemplified in the intermediate table described above, since the number of key items belonging to each category is counted in advance, the counted number of key items for each category is displayed in the GUI window for search. (For example, in FIG. 3, the number of key items for each category is displayed in the “key category list” box 172 and some other boxes), so that when the user drills down the search result, It is easy to determine which category search key is appropriate in which order.

  However, in this “column position type” nomenclature, a direct fuzzy search (a wild card can be set for a part of each search key on the search file name) is not possible. However, this problem can be solved by performing another process such as the following before the search file name creation process. That is, in the separate process, a key item that matches the specified ambiguous search key is searched from all the key items defined in the search method definition file 160, and as a result, the key items searched for are searched. The search key to be included in the search file name is determined.

  The above is the “column position type” nomenclature, which is the first of the “table correspondence type” nomenclature. Next, the “matrix position type” nomenclature, which is the second of the “table correspondence type” nomenclature, will be described.

  In the “matrix position type” nomenclature, the same process as the “column position type” nomenclature described above can be applied to create the intermediate table. The difference between the “matrix position type” method and the “column position type” method is in the process of creating the file name of the search file group. That is, in the “matrix position type” nomenclature, the code corresponding to each key item in the file name is not only the column position of each key item in the intermediate table, but also a pair of row position and column position (that is, matrix Coordinate). Since the corresponding key item can be uniquely identified from the value of the code incorporated in the file name, when the key item of a certain category is useless, the code corresponding to the useless category is the file name. (In short, it is not necessary to incorporate codes corresponding to all categories defined in the search method definition file 160 into the file name).

  A specific example is shown. According to the “matrix position type” nomenclature described above, the file name of the search file group is as follows.

<Example of file name by "column position type" naming method>
01_1_1_1_1_1_1_1_1_1.txt
02_1_2_1_2_2_2_2_1_1.txt
03_2_3_2_3_3_1_3_2_1.txt
04_3_4_1_4_4_1_4_3_1.txt
05_4_5_2_5_5_3_5_4_2.txt
06_5_6_2_6_6_4_3_5_3.txt
07_6_7_2_7_4_5_5_6_4.txt

  On the other hand, according to the “matrix position type” nomenclature, the file names of the search file group are as follows.

<Example of file name by "matrix position type"nomenclature>
01_1A_2A_3A_4A_5A_6A_7A_8A_9A_Z.txt
02_1A_2C_3A_4D_5F_6D_7D_8A_9A_Z.txt
03_1D_2E_3C_4G_5K_6A_7G_8I_9A_Z.txt
04_1G_2H_3A_4J_5P_6A_7J_8Q_9A_Z.txt
05_1J_2K_3C_4M_5U_6G_7M_8Y_9E_Z.txt
06_1M_2N_3C_4P_5Z_6J_7G_8BG_9I_Z.txt
07_1P_2R_3C_4S_5P_6M_7M_8BO_9N_Z.txt

  The code corresponding to each key item in each file name consists of a pair of numbers and alphabets. The number in each pair means the row position of the category to which each key item belongs on the intermediate table (in short, the category identification code), and the next alphabetical character indicates the column position of each key item on the intermediate table ( In short, it means a key item identification code in each category. As the alphabet, for example, alphabetic characters (26 characters) representing 26-digit numbers are used. For example, “A” indicates “1”, “B” indicates “2”,..., “AA” indicates “27”.

  In the case of the following intermediate table, if the code corresponding to a certain key item is “2R”, for example, the number “2” is (the “number” in the top row is a category not used as a search key). Since this is ignored, it refers to the category “name” in the second row, and the alphabet “R” is the 18th column in the category “name” (within the column range in which the key item is registered). The key item "Kumiko" starting from Therefore, the symbol “2R” indicates the key item “Kumiko” of the category “Name”.

<Example of intermediate table>
Number (7), 01,02,03,04,05,06,07
Mr. (6), Shoji, Castle, Otsuka, Tanaka, Okamoto, Itona (7), Wataru, Satoru, Yoshie, Hiroaki, Noriko, Yukiko, Kumiko Sex (2), Male, Female Age (7), 57, 56,47,58,40,50,59
Birth year (6), 1950, 1952, 1960, 1949, 1966, 1967
Birth month (5), 04,01,07,09,10
Birthday (5), 27,20,25,04,13
Postal code (6), 1620067,1401234,1500001,2220005,3130088,2220000
Prefectures (4), Tokyo, Aichi, Kanagawa, Saitama

In this case, the search file name is created by using a code indicating the matrix position of the specified search key in the intermediate table and a wild card as a substitute for the unspecified search key, and using the delimiter "_" It is a method of connecting via. In this method, since the category of the designated search key can be identified from the code, one wild card “*” can be set for a plurality of undesignated search key chunks. In addition, a wild card “*” is always added to both ends of the search file name. For example, when the search key “male” of the category “gender” is specified, the search file name is
* 3A *
And it is sufficient. Also, if the search key “male” for the category “sex” and the search key “Shoji” for the category “Mr.” are specified, the search file name will be
* 1A * 3A *
And it is sufficient.

  The reason for adding a wildcard "*" to both ends of the search file name is that if the wildcard "*" is not added before the first (leftmost) search key or after the last (leftmost), it matches the search key. This is because, when it is determined whether or not the character string is in the file name, it is necessary to check whether the search key is the first or last, so that it takes time to determine. Therefore, as shown in the example of the file name of the search file group described above, the suffix “_Z” is added to the end of the key item at the end (right end) of the file name of each search file. Has been. By including the suffix “_Z”, a wild card “*” can be added to the end of the search file name. Note that a key item number that is not used as a search key is attached to the beginning of the search file name, and this key item number plays the same role as the suffix “_Z” as a prefix. You can also add a wild card "*" at the beginning of.

  Even when this “matrix position type” nomenclature is adopted, the search process described with reference to FIGS. 3 to 7 can be applied as it is.

According to this “matrix position type” nomenclature, the file name can be shortened, or the number of search keys that can be incorporated into a file name with a limited length can be increased. In particular, the fact that symbols corresponding to useless categories can be omitted in the file name contributes to shortening the file name. For example, if there are 20 key item categories in total, the following file name
00032_1A_6IJ_7O_8CQ_14BT_15BT_Z.txt
In this case, the key items of the 2, 3, 4, 5, 9-13, and 16-20th categories are omitted.

  Further, since the 26-digit character incorporated in the file name indicates the position of the first character of the corresponding key item, the time taken to create the file name can be shortened. This is because the processing time is shorter when the position of a certain character in the character string is examined than when the certain item is located on a two-dimensional matrix such as an intermediate table. Even if a key item is added later, there is no need to change the existing file name. In addition, 26-digit characters are useful for encryption. For example, if a 26-digit number is encrypted using a table in which the alphabetical order is changed, it is necessary to find a combination of 26 factorials in order to find a correct answer, so that it is difficult to decrypt the code.

  So far, several improved file naming schemes have been described. In the above description, one search file is created for one data item regardless of the nomenclature. However, a plurality of search files can be created for one data item. As a result, the problem due to the limitation of the file name length can be solved. Again, any of the various nomenclatures described above are applicable.

  FIG. 8 is a data flow diagram in a modified embodiment of the present invention configured to create a plurality of search files for one data item.

  Also in this modified embodiment, the basic configuration of the system is the same as that shown in FIG. As shown in FIG. 8, the converter 104 analyzes the input search method definition file, creates a plurality of (for example, two) search subfiles 300 and 302 for each data item, and the plurality of these subfiles. Are stored in different predetermined directories.

  In the file names of the plurality of search subfiles 300 and 302 associated with one data item, key items of different category groups possessed by the data item are respectively incorporated. For example, in the example shown in FIG. 8, the key items of the first category # 1 to # 3 are incorporated in the file name of one search subfile 300, and the file name of the other search subfile 302 is included in the file name. Includes key items of categories # 4 to # 6 in the latter half. In this way, the key items of many categories possessed by each data item are divided into a plurality of groups, and the key items of different category groups are incorporated in the file names of the different search subfiles 300 and 302, respectively. It should be noted that the data item numbers of the data items associated with the file names of any of the search subfiles 300 and 302 are incorporated in common.

  The search engine 102 creates a plurality of search subfile names 304 and 306 using the specified search key, and searches the files from the different directories using the search subfile names 304 and 306, respectively. . The plurality of search subfile names 304 incorporate search keys belonging to the different category groups. For example, in the example shown in FIG. 8, one search subfile name 304 includes the search keys of the first category # 1 to # 3, and the other search subfile name 306 includes the second category #. Search keys 4 to # 6 are incorporated. If a search key belonging to a category group to be incorporated into a certain search subfile name is not specified, all search keys in the search subfile name are replaced with wild cards.

  When a plurality of sub file name lists 308 and 310 as sub search results from the different directories are returned from the file system 110, the search engine 102 synthesizes the plurality of sub file name lists 308 and 310 by a logical operation. Thus, the composite search result 312 is created. That is, if all of the plurality of subfile name lists 308 and 310 include subfile names having a common data item number, those subfile names have been hit (that is, the common data item). Means that the data item identified by the number has been hit), so that the common data item number (and the key item embedded in the subfile name having the common data item number may be added) Can be the combined search result 312.

  The preferred embodiment of the present invention and its modifications have been described above. In the above embodiment, the file name search is used, so the search speed is high. Also, the user can arbitrarily determine the order in which the plurality of search keys are specified when drilling down the search results.

  On the other hand, in the above embodiment, it is necessary to save a large number of search files in a predetermined directory in the storage device. In addition, the number of search key categories that can be used is limited by the limit on the length of file names that can be handled by a computer.

  A second embodiment of the present invention for overcoming such a problem will be described below.

  FIG. 9 shows the overall configuration of the search system according to the second embodiment. 9, blocks having the same or similar role as the configuration shown in FIG. 1 are denoted by the same reference numerals, and redundant description is omitted.

  Hereinafter, the configuration and operation of the search system according to this embodiment will be described with specific examples.

  As shown in FIG. 9, the converter 104 creates an intermediate table file 320 and a key item list file 322 based on the input search method definition file 160 (for example, a CSV format file) and saves it in the storage device 120. (Step 400).

  A specific example is shown. Assume that the key item is defined as follows by the search method definition file 160.

<Key item definition example>
Number, name, gender, age, year of birth, month of birth, date of birth, postal code, prefecture
01, Shoji, Wataru, Male, 57, 1950, 04, 27, 1620067, Tokyo
02, Shoji, Satoru, Male, 56, 1952, 01, 20, 1620067, Tokyo
03, Castle, Rye, Woman, 47, 1960, 04, 25, 1401234, Tokyo
04, Otsuka, Hiroaki, Male, 58, 1949, 04, 04, 1500001, Tokyo
05, Tanaka, Noriko, Woman, 40, 1966, 07, 13, 2220005, Aichi
06, Okamoto, Yukiko, Woman, 50, 1967, 09, 25, 3130088, Kanagawa Prefecture
07, Ito, Kumiko, Woman, 59, 1949, 10, 13, 2220000, Saitama

  In this case, the intermediate table file 320 is created in the following format, for example, in the form of a CSV file. This content is the same as the intermediate table described with respect to the nomenclature of “table correspondence type” (“column position type” and “matrix position type”) in the first embodiment described above.

<Example of intermediate table file 320>
Number (7), 01,02,03,04,05,06,07
Mr. (6), Shoji, Castle, Otsuka, Tanaka, Okamoto, Itona (7), Wataru, Satoru, Yoshie, Hiroaki, Noriko, Yukiko, Kumiko Sex (2), Male, Female Age (7), 57, 56,47,58,40,50,59
Birth year (6), 1950, 1952, 1960, 1949, 1966, 1967
Birth month (5), 04,01,07,09,10
Birthday (5), 27,20,25,04,13
Postal code (6), 1620067,1401234,1500001,2220005,3130088,2220000
Prefectures (4), Tokyo, Aichi, Kanagawa, Saitama

  Next, based on the contents of the intermediate table file 320, a key item list file 322 is created with the following contents, for example, in the form of a text file.

<Example of key item list file 322>
01_1A_2A_3A_4A_5A_6A_7A_8A_9A_Z
02_1A_2C_3A_4D_5F_6D_7D_8A_9A_Z
03_1D_2E_3C_4G_5K_6A_7G_8I_9A_Z
04_1G_2H_3A_4J_5P_6A_7J_8Q_9A_Z
05_1J_2K_3C_4M_5U_6G_7M_8Y_9E_Z
06_1M_2N_3C_4P_5Z_6J_7G_8BG_9I_Z
07_1P_2R_3C_4S_5P_6M_7M_8BO_9N_Z

  As described above, the key item list file 322 is the file name of each search file described with respect to the naming method of “table correspondence type” (“column position type” and “matrix position type”) in the first embodiment. Each character string obtained by removing the extension from (that is, each character string formed by concatenating codes indicating matrix positions on the intermediate table file 320 of a plurality of key items of each data item) It is nothing but what was written in one line of the text file (in other words, distinguished from each other). Therefore, the key item list file 322 describes a large number of the character strings (character string mother groups) associated with a large number of data items. Since the search using the search key is performed on the key item list file 322, the number of search keys that can be specified is not limited by the length of the file name.

  Each time the search engine 102 receives a search request (step 130), the search engine 102 reads the intermediate table file 320 and the key item list file 322, and stores the contents of the files 320 and 322, that is, the intermediate table 324 and the key item list 326, on the computer. Is expanded on the main memory 328 (step 402). A search for a data item using the search key (step 404) is performed with reference to the intermediate table 324 and the key item list 326 on the main memory 328. There is no problem that the number of search keys that can be used like the file name search is limited by the length of the file name. The intermediate table 324 is a character string for searching the designated search key in the key item list 326 (that is, a code indicating the matrix position of the search key on the intermediate table file 320, in other words, the search key). Key identification code).

  In the data item search using the search key (step 404), since it is not a file name search, a search using the wild card “*” cannot be performed. Instead, a function for checking whether a certain character string “YYY” exists in a certain character string “XXXX” is used. For example, when a search is performed with the search key “male” of the category “gender” and the search key “Tokyo” of the category “prefecture”, the search is performed for each row of the key item list 326 using the function. A process that checks whether or not the character string “_3A_” corresponding to the key “male” is included and the character string “_9A_” corresponding to the search key “Tokyo” is included, and the first number of the hit line is returned. Execute.

  When a plurality of search keys are arranged in succession, for example, when searching with the search key “Wataru” of the category “name” and the search key “male” of the adjacent category “gender”, each row of the key item list 326 is displayed. On the other hand, it is only necessary to check whether or not the character string “_2A_3A_” concatenated with codes corresponding to the consecutive search keys is included.

  Also in the second embodiment, the search process described with reference to FIGS. 3 to 7 can be basically applied. However, in the search process described with reference to FIGS. 3 to 7, the file name of the search file that matches the search file name incorporating the specified search key is searched, whereas this second implementation is performed. The search process in the form includes a character string (that is, a code indicating the matrix position of the search key on the intermediate table file 320, in other words, an identification code of the search key) corresponding to the specified search key. The character string in the line on the key item list file 322 is searched. In processes other than this difference, the search process described with reference to FIGS. 3 to 7 can be applied.

  According to the second embodiment described above, the number of search keys that can be used is not limited. The selection of search keys and the order of search keys can be freely determined by the user each time, and this does not affect the search speed. Regardless of the number of data items, two files, the intermediate table file 320 and the key item list file 322, are sufficient for the number of files to be stored in the storage device, so that database maintenance and backup are extremely easy.

  In the second embodiment, the “table correspondence type”, particularly “matrix position type” nomenclature described above is used as a method of creating a character string associated with each data item in the key item list file 322. However, other nomenclatures already described, such as “direct type”, “encoding type” or “column position type” nomenclature can be used instead (“direct type” or When using the “encoding type” nomenclature, the intermediate table file 320 is not required). In the second embodiment, one key item list file 322 is used. Instead, a plurality of subfiles having contents equivalent to those obtained by dividing the contents of the key item list file 322 are used. You can also.

  The several embodiments of the present invention described above are examples for explaining the present invention, and are not intended to limit the scope of the present invention. The present invention can be implemented in various forms different from these embodiments without departing from the gist thereof.

1 is a block diagram showing an overall configuration of a search system according to a first embodiment of the present invention. The data flow figure which shows the specific example of the production method of the search files 122,122, ... by the converter 104. FIG. The figure which shows the example of the graphical user interface which the search engine 102 provides. Part of a process flow diagram of a search performed by the search engine 102. Part of a process flow diagram of a search performed by the search engine 102. Part of a process flow diagram of a search performed by the search engine 102. Part of a process flow diagram of a search performed by the search engine 102. The data flow figure of the deformation | transformation embodiment of this invention comprised so that the some file for a search might be produced with respect to one data item. The block diagram which shows the whole structure of the search system concerning the 2nd Embodiment of this invention.

Explanation of symbols

100 Search System 102 Search Engine 104 Converter 110 OS File System 120 Storage Device 122 Search File 160 Search Method Definition Information 170 GUI Window 300, 302 Search Subfile 304, 306 Search Subfile Name 312 Composite Search Result 320 Intermediate Table File 322 Key item list file 324 Intermediate table 326 Key item list

Claims (11)

In a system for searching for a data item that matches a search key from a number of data items,
A storage means for storing a character string group composed of a large number of character strings respectively associated with a large number of data items, wherein each character string in the character string group belongs to a plurality of categories related to each data item. Including the concatenation of a plurality of key items or the concatenation of identification codes of the plurality of key items;
A first search key selecting means for selecting a search key from the key items included in the character string population according to the user's selection;
Search means for issuing a search request for searching the character string group for a character string that matches the selected search key;
Key item display means for extracting a key item belonging to the category specified by the user from the searched character string obtained as a search result for the search request, and displaying the extracted key item;
A next search key is selected from the displayed key items according to the user's selection, and the selected next search key and the search key selected by the first search key selection step are selected. Second search key selection means for operating the search means again so as to issue a search request for searching for a character string matching both from the character string population;
The second search key selection means repeatedly operates a plurality of times in accordance with the selection of the user, and each time, the next search key selected this time, the search key selected in the previous time, and the first search key are selected. For data retrieval, the retrieval means is operated again so as to issue a retrieval request for retrieving from the character string group a character string that matches all the retrieval keys selected by the retrieval key selection step. System. In a method for searching for a data item that matches a search key from a number of data items,
A storage step of storing a character string group composed of a number of character strings respectively associated with a number of data items, wherein each character string in the character string group belongs to a plurality of categories related to each data item Including the concatenation of a plurality of key items or the concatenation of identification codes of the plurality of key items;
A first search key selection step of selecting a search key from the key items included in the character string group according to a user's selection;
A search step for issuing a search request for searching the character string group for a character string that matches the selected search key;
A key list display step of extracting and displaying key items belonging to the category designated by the user from the searched character strings obtained as search results for the search request;
A next search key is selected from the displayed key items according to the user's selection, and the selected next search key and the search key selected by the first search key selection means are selected. A second search key selection step for causing the search means to operate again so as to issue a search request for searching for a character string matching both from the character string group;
A second search key selection step as means;
The second search key selection step is repeated a plurality of times in accordance with repetition of the user's selection, each time, the next search key selected this time, the search key selected in the previous time, and the first search key selection A repetitive step of causing the search means to operate again so as to issue a search request for searching the character string group for a character string that matches all of the search keys selected by the means. Way for. In a computer program for causing a computer to execute a method for searching for a data item that matches a search key from a number of data items, the method includes:
A storage step of storing a character string group composed of a number of character strings respectively associated with a number of data items, wherein each character string in the character string group belongs to a plurality of categories related to each data item Including the concatenation of a plurality of key items or the concatenation of identification codes of the plurality of key items;
A first search key selection step of selecting a search key from the key items included in the character string group according to a user's selection;
A search step for issuing a search request for searching the character string group for a character string that matches the selected search key;
A key list display step of extracting and displaying key items belonging to the category designated by the user from the searched character strings obtained as search results for the search request;
A next search key is selected from the displayed key items according to the user's selection, and the selected next search key and the search key selected by the first search key selection means are selected. A second search key selection step for causing the search means to operate again so as to issue a search request for searching for a character string matching both from the character string group;
A second search key selection step as means;
The second search key selection step is repeated a plurality of times in accordance with repetition of the user's selection, each time, the next search key selected this time, the search key selected in the previous time, and the first search key selection A data search comprising: repeating the search means to issue a search request for searching the character string group for a character string that matches all of the search keys selected by the means. Computer program for. The computer program according to claim 3, wherein
The storing step includes storing a number of search files associated with the number of data items;
A computer program for data retrieval, wherein each of the plurality of character strings is incorporated in a file name of each of the plurality of stored search files. The computer program according to claim 4, wherein
In each of the search steps,
As a substitute for the next search key selected this time, the search key selected in the previous round, the search key selected by the first search key selection means, and an unselected search key of another category Creating a search file name by concatenating wildcards;
Issuing a search request for searching for a file name of the search file that matches the created search file name. The computer program according to claim 5, wherein
In each character string incorporated in the file name of each search file, each key item or each key item identification code is sandwiched from both sides by a predetermined delimiter so that the plurality of key items or the plurality of key items A data search in which identification codes of key items are connected and a predetermined prefix and suffix are added to the beginning and end of the plurality of key items or a combination of identification codes of the plurality of key items, respectively. Computer program for. In the computer program as described in any one of Claims 4-6,
A computer program for data search, in which each search file stores reference information of each data item. The computer program according to claim 3, wherein
The storing step includes a step of storing a key item list file;
A computer program for data retrieval, wherein the plurality of character strings are distinguished from each other and stored in the stored key item list file. In the computer program as described in any one of Claims 3-8,
Each of the character strings includes a concatenation of identification codes of the plurality of key items related to the data items,
The identification code of each key item is obtained by converting the character code of each of the characters constituting each key item into a character string representing an M-ary number (M is a number greater than 2). program. In the computer program as described in any one of Claims 3-8,
The method further comprises the step of saving an intermediate table file in which a number of key items relating to the number of data items are registered;
Each of the character strings includes a concatenation of identification codes of the plurality of key items related to the data items,
A computer program for data retrieval, wherein an identification code of each key item is a code for identifying each key item on the intermediate table file. The computer program according to claim 10, wherein
A computer program for data search, wherein the identification code of each key item is a code for specifying the position of each key item on the intermediate table file.

Images