JP2009048352A - Information retrieval device, information retrieval method and information retrieval program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information retrieval device capable of executing retrieval processing at high speed without consuming a lot of resources. <P>SOLUTION: This information retrieval device 1 has: a retrieval condition acquisition means 10 for acquiring a retrieval condition; an index acquisition means 13 for acquiring a bitmap table from a topmost stage to a downmost stage in turns based on the retrieval condition from an index wherein the bitmap tables each hierarchically formed by degenerating or compressing the bitmap table are layered in a multistage state; a decompression means 14 for decompressing data of the compressed bitmap table when the bitmap table is compressed; a bitmap calculation means 15 for extracting a bitmap string corresponding to the retrieval condition from the bitmap table to the downmost stage in turns in the decompressed bitmap table, and performing a bitwise operation to calculate a bitmap; and a bitmap correspondence information extraction means 12 for extracting a document corresponding to a position allocated with a bit from a plurality of documents in the bitmap calculated in the downmost stage. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information search device for searching for search target information desired by a user from a large amount of search target information such as a large number of document files, using a word included in the search target information and attribute information of the search target information as search conditions The present invention relates to an information search method and an information search program.

  Conventionally, for example, from a large amount of search target information (text information) such as a large number of document files, search using “words included in search target information” or “attribute information of search target information” as search condition information The following two methods are known as methods for retrieving target information (for example, see Non-Patent Document 1 and Non-Patent Document 2). For example, if the search target information is a document file, the attribute information of the search target information indicates a file creation date and time, a creator, the number of constituent characters, browseability information, an expiration date, and the like.

  In the first conventional search method, a plurality of search conditions (words included in the search target information and attribute information of the search target information) are separated into individual search conditions in relation to the method disclosed in Non-Patent Document 1. Thus, search processing is performed on a large amount of search target information using each separated search condition independently, and a primary search result set is obtained for each search condition. Specifically, when, for example, “a word included in the search target” and “one attribute information of the search target information” are input to the conventional information search device as the search conditions, the information search device First, as a first step, all search target information including the input “words included in the search target” is searched from a large amount of search target information, and a plurality of search targets that are primary search results Create and retain information as a search result set. Next, as a second stage, the information search apparatus, for each piece of search target information included in the primary search result set, satisfies the text that satisfies “one attribute information of the search target information” input as a search condition. Information is extracted and presented to the user as a final search result. Thereby, the user can acquire desired text information according to the input “words included in the search target” and “one attribute information of the search target information”.

In addition, the conventional second search method corresponds to all search conditions that may be inquired for a large amount of search target information in relation to the methods disclosed in Non-Patent Document 1 and Non-Patent Document 2. A bitmap index created in advance is prepared. In this method, when a plurality of search conditions are input to the information search apparatus, a logical product between bitmap sequences corresponding to each search condition is calculated based on a bitmap index created in advance, and a bit of the logical product result Information corresponding to the place where the user stands is presented as information desired by the user.
J. Zobel and A. Moffat, Inverted Files Text Search Engines, ACM Computing Surveys, Vo1.38, No. 2, Article 6, July, 2006 CY Chan and YE Ioannidis, Bitmap Index Design and Evaluation, Proceedings of the 1998 ACM SIGMOD international Conference on Management of data, pages: 355-366, 1998.

  However, in the first conventional search method, a plurality of input search conditions are separated into individual search conditions, and only one of the plurality of input search conditions is used, and other search conditions are set. Acquire a primary search result set without using it. Therefore, there is a possibility that the number of search targets included in this search result set will be large. In this case, a large amount of resources (mainly memory or disk) are consumed to store the search result set. In addition, the first conventional search method includes an individual stage for acquiring a primary search result set corresponding to each search condition separated into individual search conditions, and a final presentation to the user. There is a problem that it takes time for the search process.

  In the second conventional search method, the size of the bitmap index corresponding to all search conditions that may be inquired when dealing with, for example, 10 million or more document files as a large amount of search target information. Becomes huge and requires a lot of resources (mainly memory or disk) to maintain it. Furthermore, when performing bitwise calculations, a huge bit string based on a bitmap index of a huge size must be handled, which consumes a lot of resources (mainly CPU cache and memory) and degrades the calculation performance. There is a problem.

  Accordingly, an object of the present invention is to provide an information search apparatus, an information search method, and an information search program capable of solving the above-described problems and performing search processing at high speed without consuming a large amount of resources. To do.

  In order to solve the above problem, the information search device according to claim 1 is a search target corresponding to the input search condition from a plurality of search target information based on the search condition input by the user. An information search device for searching for information as search target information desired by the user, wherein a plurality of predetermined search conditions for the search target information desired by the user are acquired as the search conditions. From the index that is configured by sequentially stacking a plurality of bitmap tables that are processed by performing a process of reducing or compressing the predetermined bit data stored in the condition acquisition means and the bitmap table created in advance, Based on a plurality of search conditions acquired by the search condition acquisition means, a bitmap table is acquired in order from the largest to the smallest data inclusion relationship. Index acquisition means; and decompression means for generating a bitmap table in which the compressed data is decompressed when predetermined data stored in the bitmap table acquired by the index acquisition means is compressed, and In the bitmap table acquired by the index acquisition means and decompressed when the predetermined data is compressed, the inclusion relation of the data from the largest to the smallest in order, A plurality of bitmap sequences corresponding to a plurality of search conditions acquired by the search condition acquisition unit are extracted from the bitmap table, and assigned to corresponding positions of the extracted plurality of bitmap sequences. Bitmap calculation means for calculating a bitmap by performing a bitwise operation on the bit, and the data As the information corresponding to the bitmap calculated for the bitmap table having the smallest inclusion relationship, the one bit is assigned in the bitmap calculated for the minimum bitmap table from among the plurality of search target information. Bit map correspondence information extracting means for extracting search target information corresponding to the bit position, and the previously created bitmap table includes identification information of the plurality of search target information, and the plurality of predetermined The search condition is associated with one of the binary bits, and the predetermined data stored in the pre-process bitmap table is included in the post-process bitmap table. It is characterized by being hierarchically formed.

  According to such a configuration, the information search apparatus supports a search condition from an index configured by stacking a plurality of bitmap tables formed hierarchically based on a bitmap table created in advance and configured in a multistage manner. A plurality of bitmap tables are obtained sequentially from the upper layer. Here, in the bitmap table, the plurality of predetermined search conditions associated with the identification information of the plurality of search target information are, for example, words that can be input as search conditions included in the desired search target information. Or attribute information of desired search target information. Then, the information retrieval apparatus decompresses the compressed data in the acquired plurality of bitmap tables if there is compressed data, and sequentially searches the plurality of bitmap tables from the target bitmap table as processing targets from the upper layer. A bitmap is calculated by extracting a plurality of bitmap sequences corresponding to the conditions and performing a bitwise operation. In the bit map calculated in the process of this sequential processing, data necessary for calculation can be extracted from the bit map of the lower layer to be processed using the position where the bit is set as an index. Finally, the identification information of the search target information corresponding to the position where the bit is set is extracted as the search target information to be presented to the user in the bitmap calculated from the lowest bit map. Therefore, the search processing apparatus can output the extracted search target information as search result information corresponding to the search condition input by the user.

  As a result, the information search apparatus does not need to prepare in advance a bitmap index corresponding to all search conditions that may be inquired about a large amount of search target information. Memory and disk) are not consumed. In addition, the information search apparatus can handle a plurality of search conditions at once rather than separately processing them, and can narrow down search candidates at high speed. Therefore, it is possible to reduce the consumption of resources (disk, memory, CPU) as compared with the conventional apparatus and to execute the search process at high speed.

  Further, the information search device according to claim 2 is the information search device according to claim 1, wherein the index uses the group as a unit when the plurality of search target information is divided into a plurality of groups. It is configured in a multi-stage manner by sequentially performing a process of reducing or compressing predetermined data stored in a bitmap table created in advance, and the bitmap table showing each stage of the index is based on a bit-wise operation A join process is possible.

  According to this configuration, the information search apparatus acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group of search target information in the index. According to this, the information search apparatus searches only the bitmap string related to the group of search target information corresponding to the input search condition from the bitmap table to be processed. Here, the group of search target information can be provided hierarchically.

  In addition, since an index that enables join processing by bitwise calculation is used in each stage, the following effects can be obtained. First, when there is compressed data at each stage in the index, only a necessary part needs to be partially decompressed, so that high-speed search is possible. Further, the intermediate data size can be reduced during the search process. Furthermore, the bitmap size in one calculation can be reduced, and as a result, the CPU cache can be used efficiently. Therefore, the information search apparatus having such a configuration can reduce resource consumption and execute the search process at a higher speed than the conventional apparatus.

  The information search device according to claim 3 is the information search device according to claim 1, wherein the index divides the plurality of search target information into a plurality of groups based on the similarity of contents. A bit map indicating each level of the index, wherein the group is configured in a multistage manner by sequentially performing a process of reducing or compressing predetermined data stored in the previously generated bitmap table in units of the group. The table is characterized in that join processing by bitwise operation is possible.

  According to this configuration, the information search apparatus acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group whose contents are similar with respect to search target information in the index. Here, groups similar in content to search target information can be provided hierarchically. The similarity can be determined as appropriate according to the search target information. For example, if the search target information is a document, it may be determined by the title and genre of the document, or may be determined by the keyword included in the document and the number of keywords. Moreover, according to the information search device having such a configuration, the search target information is divided into groups having similar contents, and the following effects are obtained. That is, since the places where the bits stand in the index are dense, the compression efficiency is improved. In addition, the number of places where the bit stands after degeneration is reduced. As a result, the number of bits standing in the bitmap operation result is reduced, so that consumption of resources (memory, disk) can be suppressed.

  The information search device according to claim 4 is the information search device according to any one of claims 1 to 3, wherein the index includes the plurality of predetermined search conditions as a plurality of groups. When the data is divided into groups, the group is configured in a multi-stage manner by sequentially performing a process of reducing or compressing the predetermined data stored in the bitmap table created in advance, and each index of the index The bit map table indicating the stages is characterized in that join processing by bit-wise calculation is possible.

  According to this configuration, the information search apparatus acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group of search conditions in the index. Here, groups of search conditions can be provided hierarchically. According to this, the information retrieval apparatus retrieves only the bitmap string related to the group of retrieval conditions in the bitmap table from the bitmap table to be processed according to the inputted retrieval condition. Therefore, it is possible to reduce resource consumption and execute the search process at a higher speed than the conventional apparatus.

  Further, the information search device according to claim 5 is the information search device according to any one of claims 1 to 3, wherein the index has the plurality of predetermined search conditions similar in content. When the data is divided into a plurality of groups based on the degree, the group is configured in multiple stages by sequentially performing a process of reducing or compressing predetermined data stored in the previously created bitmap table in units of the group The bit map table showing each stage of the index can be joined by a bitwise operation.

  According to such a configuration, the information search apparatus acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group whose contents are similar with respect to search conditions in the index. Here, groups similar in content to the search conditions can be provided hierarchically. The similarity can be determined as appropriate according to the search condition. For example, if the search condition is a keyword included in a document, it may be determined by a higher-level concept classification of whether the keyword is related to food or location. If the search condition is the document creation date and time, the date may be determined. Moreover, according to the information search device having such a configuration, the following operational effects can be obtained by dividing the search conditions in the index into groups having similar contents. That is, since the places where the bits stand in the index are dense, the compression efficiency is improved. In addition, the number of places where the bit stands after degeneration is reduced. As a result, the number of bit-settled portions in the bitmap operation result is reduced, so that consumption of resources (memory, disk) can be suppressed.

  The information search device according to claim 6 is the information search device according to any one of claims 1 to 5, wherein the plurality of predetermined search conditions in the index are a predetermined sequence. A plurality of sections configured by dividing values, information obtained by digitizing information related to the plurality of search target information is included in any of the plurality of sections, and the index includes the section As a unit, the predetermined data stored in the bitmap table prepared in advance is subjected to a process of reducing or compressing sequentially, and is configured in a multistage shape.

  According to such a configuration, the information search device acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned for each section in the index. Here, the information obtained by digitizing the information related to the search target information is, for example, the price, the number of characters, the number of pages, etc. of the search target information. According to this, the information search device corresponds to the section corresponding to the information (for example, price) obtained by digitizing information related to the search target information as the input search condition from the bitmap table to be processed. Only the bitmap sequence related to is searched. Therefore, it is possible to reduce resource consumption and execute the search process at a higher speed than the conventional apparatus.

  Moreover, in order to solve the said subject, the information search method of Claim 7 respond | corresponded to the said input search condition from several search object information based on the search condition input by the user. An information search method for an information search device for searching for search target information as search target information desired by the user, wherein the information search device is configured to search for the search target information desired by the user as the search condition. A search condition acquisition step for acquiring a plurality of predetermined search conditions, and a plurality of bitmap tables processed by performing a process of reducing or compressing predetermined bit data stored in a bitmap table created in advance From among the indexes that are sequentially stacked in a shape, the data inclusion relationship is the maximum based on the plurality of search conditions acquired in the search condition acquisition step. An index acquisition step for acquiring a bitmap table in order from 1 to the minimum, and when the predetermined data stored in the bitmap table acquired in the index acquisition step is compressed, the compressed data is A decompression step for generating a decompressed bitmap table, and a bitmap table obtained by the index acquisition step and decompressed when the predetermined data is compressed, the compressed data being decompressed. In order from the largest to the smallest inclusive relationship, a plurality of bitmap sequences corresponding to the plurality of search conditions acquired in the search condition acquisition step are extracted from the bitmap table, and the extracted plurality Bitwise operation on the bits assigned to the corresponding positions in the bitmap sequence of A bitmap calculation step for calculating a bitmap by performing, and information corresponding to the bitmap calculated for the bitmap table with the smallest data inclusion relationship, from among the plurality of search target information, the minimum A bitmap correspondence information extraction step of extracting search target information corresponding to a bit position to which the one bit is assigned in the bitmap calculated for the bitmap table, and the bitmap table created in advance is The identification information of the plurality of search target information and the plurality of predetermined search conditions are associated with any one of the binary bits, and the index is stored in the bitmap table before the processing The predetermined data is hierarchically formed so as to be included in the bitmap table after the processing. Features.

  According to such a procedure, the information search apparatus can cope with a search condition from an index configured by stacking a plurality of bitmap tables formed hierarchically based on a bitmap table created in advance. A plurality of bitmap tables are obtained sequentially from the upper layer. Then, the information retrieval apparatus decompresses the compressed data in the acquired plurality of bitmap tables if there is compressed data, and sequentially searches the plurality of bitmap tables from the target bitmap table as processing targets from the upper layer. A bitmap is calculated by extracting a plurality of bitmap sequences corresponding to the conditions and performing a bitwise operation. Then, the search processing apparatus finally searches the user for the identification information of the search target information corresponding to the position where the bit is set in the bitmap calculated from the lowest bit map. Extract as information.

  An information search method according to claim 8 is the information search method according to claim 7, wherein the index divides the plurality of search target information into a plurality of groups based on similarity of contents. A bit map indicating each level of the index, wherein the group is configured in a multistage manner by sequentially performing a process of reducing or compressing predetermined data stored in the previously generated bitmap table in units of the group. The table is characterized in that join processing by bitwise operation is possible.

  According to this procedure, the information search apparatus acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group whose contents are similar with respect to search target information in the index. Here, the similarity can be appropriately determined according to the search target information.

  An information search method according to claim 9 is the information search method according to claim 7 or claim 8, wherein the index includes a plurality of predetermined search conditions based on similarity of contents. The index is configured in a multi-stage manner by sequentially performing a process of reducing or compressing predetermined data stored in the previously created bitmap table in units of the group when divided into groups. The bit map table showing each stage of the above is characterized in that join processing by bitwise calculation is possible.

  According to such a procedure, the information search device acquires a plurality of bitmap sequences based on a bitmap table in which bits are assigned to each group whose contents are similar with respect to search conditions in the index. Here, the similarity can be appropriately determined according to the search condition.

  The information search program according to claim 10 is an information search program for realizing the function of the information search device according to any one of claims 1 to 6 by a computer. By being configured in this way, a computer in which this program is installed can realize each function based on this program.

  According to the present invention, it is not necessary to prepare in advance a bitmap index corresponding to all search conditions that may be inquired about a large amount of search target information, and a plurality of search conditions can be collected together without being separated individually. Since processing is performed, search processing can be executed at high speed without consuming a large amount of resources.

  The best mode for carrying out an information search apparatus and information search method of the present invention (hereinafter referred to as “embodiment”) will be described in detail below with reference to the drawings.

[Configuration of information retrieval device]
FIG. 1 is a configuration diagram schematically showing an information search apparatus according to an embodiment of the present invention. The information search device 1 searches, as search target information desired by the user, search target information corresponding to the input search condition from among a plurality of search target information based on the search condition input by the user. To do. The information retrieval apparatus 1 is composed of, for example, a central processing unit (CPU), a random access memory (RAM), a read only memory (ROM), a hard disk drive (HDD), and an input / output interface. As shown in FIG. 1, the information retrieval apparatus 1 includes an input unit 2, a storage unit 3, a control unit 4, and an output unit 5.

<Input means>
The input unit 2 includes a predetermined communication interface and input interface. The input means 2 is for inputting, for example, user search conditions received via a communication network such as the Internet to the control means 4. Note that the input unit 2 can also input to the control unit 4 a search condition input from an input device M such as a mouse or a keyboard connected to the information search device 1 used by the user with a cable.

<Storage means>
The storage unit 3 includes a ROM that stores a predetermined program, a RAM that is used for arithmetic processing by the control unit 4, and an HDD. As shown in FIG. 1, the storage unit 3 includes an index storage unit 6, an information storage unit 7, and a program storage unit 8.

≪Index storage section≫
The index storage unit 6 stores an index created in advance, and includes, for example, a general HDD. The index is obtained in advance from a bitmap table created in advance in which identification information of a plurality of pieces of search target information and a plurality of predetermined search conditions are associated with one of the binary bits (1 or 0). It has been created. This index is configured by sequentially stacking a plurality of bitmap tables processed in a multi-stage manner by performing a process of reducing or compressing predetermined bit data stored in a bitmap table created in advance. The index is hierarchically formed so that predetermined data stored in the bitmap table before processing by the reduction or compression is included in the bitmap table after processing. In this embodiment, the index is associated with a bit of “1” for a plurality of predetermined search conditions. The multi-stage configuration is a configuration that can provide a method of accessing a corresponding bit map in the lower stage or data obtained by compressing the bit map from a place where a bit is set in the upper bit map table.

  In the present embodiment, it is assumed that in the bitmap table, a word included in each document of a plurality of documents (search target information) and attribute information of the document itself are associated with a bit of “1”. Further, in this embodiment, the index is obtained when a plurality of search target information is divided into a plurality of groups, and the group is used as a unit, and a plurality of predetermined search conditions are divided into a plurality of groups. Each group is configured in multiple stages by sequentially performing a process of reducing or compressing predetermined data stored in a bitmap table created in advance, and the bitmap table indicating each stage of the index is: Assume that join processing by bitwise calculation is possible. Here, it is preferable that each group is composed of similar contents. The index may be divided into a plurality of groups or only one of a plurality of search target information and a plurality of predetermined search conditions, or may not be grouped.

  Here, an example of a method for constructing an index will be described with reference to FIGS. 2 to 7 (refer to FIG. 1 as appropriate). FIG. 2 is a flowchart illustrating an example of a method for constructing an index. This method is executed, for example, by a predetermined information processing apparatus (not shown) that stores a program for constructing an index. The information retrieval apparatus 1 may perform the same function by further executing this program.

  First, the information processing apparatus includes all information (document) stored in the information storage unit 7 (see FIG. 1) of the information search apparatus 1 and all search conditions that can be input (word and attribute information included in the document). ) Is created (step S1). An example of this huge bitmap table is shown in FIG. FIG. 3 is a diagram illustrating an example of a bitmap table from which an index is constructed. The bitmap table 30 shown in FIG. 3 includes “excursions”, “fried eggs” and the like as “words included in the document”. Further, the bitmap table 30 includes “expiration date”, “viewability”, etc. as “attribute information of the document itself”. In FIG. 3, a portion where the bit “1” is set indicates that the corresponding document satisfies the search condition. From FIG. 3, for example, a document titled “Let's go climbing” includes the words “Tokyo”, “Tent”, and “Lamp” in the document, and is “not within the expiration date”. You can see that it is not allowed to view.

  In the flowchart of FIG. 2, following step S1, the information processing apparatus divides all information (documents) stored in the information storage unit 7 (see FIG. 1) into groups that are similar in content (step S2). All search conditions that can be input are divided into groups similar in content (step S3). FIG. 4 shows an example of a bitmap table rearranged (modified) into groups with similar contents. FIG. 4 is a diagram showing an example of a bitmap table obtained by modifying the bitmap table shown in FIG. 3 by grouping. The modified bitmap table 40 illustrated in FIG. 4 includes “outdoor”, “food”, and the like as the similar group 41 of search conditions. The modified bitmap table 40 includes “travel”, “recipe”, and the like as the document similarity group 42. Further, in the modified bitmap table 40, unique identification information (Doc id) is assigned to all documents in order to identify the documents. Compared with the bitmap table 30 shown in FIG. 3, the modified bitmap table 40 shows that the places where the bits of “1” are set are partially dense.

  In the flowchart of FIG. 2, following step S3, the information processing apparatus degenerates or compresses the modified bitmap table for each group of documents and search conditions (step S4). An example of a bitmap table obtained by degenerating the modified bitmap table is shown in FIG. FIG. 5 is a diagram illustrating an example of a bitmap table obtained by degenerating the modified bitmap table illustrated in FIG. In the degenerate bitmap table 50 shown in FIG. 5, the bit “1” indicates that the bit “1” stands even at one place in the corresponding portion of the modified bitmap table 40 shown in FIG. 4. A degeneration rule of setting a bit of “1” is applied. That is, unless all the bits in the corresponding portion of the modified bitmap table 40 shown in FIG. 4 are “0”, the degenerate bitmap table 50 is not a “0” bit. In the modified bitmap table 40 shown in FIG. 4, the places where the bits are raised are partially concentrated, so that the number of places where the bits of “1” are raised is reduced in the degenerate bitmap table 50.

  More specifically, the degenerate bitmap table 50 includes “condition group 1”, “condition group 2”, and the like as similar groups 51 of search conditions. The “condition group 1” is the same as the similar group “outdoor” shown in FIG. Also, the degenerate bitmap table 50 includes “Doc group 1”, “Doc group 2”, and the like as the Doc group 52 for the document. “Doc group 1” is the same as the similar group “travel” shown in FIG. In FIG. 5, a portion where the bit “1” is set indicates that at least one of the documents included in the corresponding Doc group satisfies at least one of the search conditions included in the similar group of search conditions (search condition group). Indicates that For example, paying attention to the bit “1” indicated by reference numeral 53, any of the documents included in “Doc group 1” (documents with Doc ids “l”, “2”, “3”, and “4”). It can be seen that one of the search conditions included in “condition group 1” (condition that includes any of the words “excursion”, “tent”, and “lamp”) is satisfied. For example, when attention is paid to the bit “0” indicated by reference numeral 54, it can be seen that none of the documents included in “Doc group 1” satisfies the search condition included in “condition group 1”.

  Next, FIG. 6 shows an example of a bitmap table obtained by compressing the modified bitmap table. FIG. 6 is a diagram showing an example of a bitmap table obtained by compressing the modified bitmap table shown in FIG. The compression bitmap table 60 includes “condition group 1”, “condition group 2”, and the like as the condition group name 61 of the search condition. The “condition group 1” is the same as the similar group “outdoor” shown in FIG. The compressed bitmap table 60 includes “Doc group 1”, “Doc group 2”, and the like as the Doc group name 62 for the document. “Doc group 1” is the same as the similar group “travel” shown in FIG.

  More specifically, in the compressed bitmap table 60, for example, the compressed data 1-1 indicated by reference numeral 63 is a document (Doc id) included in the similar group “travel” in the modified bitmap table 40 shown in FIG. Is a document that includes “l”, “2”, “3”, “4”) and a search condition included in “outdoors” (a condition that includes any of the words “excursion”, “tent”, “lamp”) ) And a bitmap portion that satisfies both of these. In the modified bitmap table 40 shown in FIG. 4, the compression bit map table 60 has improved compression efficiency because the portions where the bits stand are partially concentrated. This reduces resource consumption (memory, disk). For example, in the compressed data 1-2 indicated by reference numeral 64, since all the bits of the corresponding portion in the modified bitmap table 40 shown in FIG. 4 are “0”, it can be said that even the compression work is unnecessary. .

  The reduction or compression method performed in step S4 is a reduction or compression method that retains the boolean attribute (either 0 or 1) of the bitmap to be reduced or compressed. In other words, this method is a degeneracy or compression method that allows join processing by bitwise arithmetic. As a result, join processing by bitwise calculation at each stage of the index becomes possible. In the above-mentioned example, reduction is performed while retaining the boolean attribute of the bitmap, and compression is performed without retaining the boolean attribute of the bitmap, but in some cases, reduction is not performed and the boolean attribute of the bitmap is It is also possible to perform compression while holding

  In the flowchart of FIG. 2, following step S4, the information processing apparatus configures the reduced or compressed bitmap table in a multistage manner (step S5). As an example, a two-stage index is shown in FIG. FIG. 7 is a diagram showing an example of a multi-stage (two-stage) index obtained by stacking the degenerate bitmap table shown in FIG. 5 on the compressed bitmap table shown in FIG. As shown in FIG. 7, the bit “1” indicated by reference numeral 71 a in the first stage 71 of the index corresponds to “compressed data 2-1” in the second stage (lowermost stage) 72 of the index. Similarly, the bit “1” indicated by reference numeral 71b corresponds to “compressed data 2-2”, and the bit “1” indicated by reference numeral 71c corresponds to “compressed data 2-M”.

  In the flowchart of FIG. 2, following step S5, the information processing apparatus determines whether or not the number of stages is optimal (step S6), and repeats the processes of steps S2 to S5 until the optimal number of stages is reached. As a result, an index stored in the index storage unit 6 (see FIG. 1) is constructed. Here, “optimal” indicates whether or not the overall size of the index and the bitmap and compressed data in the bitmap table at each stage are the optimum size for handling, and a predetermined standard is determined in advance. . This predetermined standard differs depending on the environment in which the document (search target information) is handled, and is determined appropriately. Also, as shown in FIGS. 5 and 6, both reduction and compression can be performed by one grouping, and therefore, when the processing of step S2 to step S5 is repeated, the group is not necessarily changed. Not exclusively. Further, when the process of compressing the predetermined data stored in the predetermined bitmap table is repeated hierarchically, the group is changed.

Returning to FIG. 1, the description of the configuration of the information search apparatus 1 will be continued.
≪Information storage section≫
The information storage unit 7 stores a plurality of pieces of text information (search target information) to be searched, and includes, for example, a general HDD. In the present embodiment, each piece of text information stored in the information storage unit 7 is a document. These documents are provided with identification information (document ID: Doc id) for identifying the document and a document title.

≪Program storage section≫
The program storage unit 8 stores various programs to be executed by the control unit 4 and includes, for example, a general memory, an HDD, and the like. The index storage unit 6 and the information storage unit 7 can also be configured with one or more external storage devices.

<Control means>
The control means 4 is composed of, for example, a CPU and includes a search condition acquisition means 10, a bitmap information calculation means 11, and a bitmap correspondence information extraction means 12. The search condition acquisition unit 10, the bitmap information calculation unit 11, and the bitmap correspondence information extraction unit 12 execute a predetermined program stored in the program storage unit 8 of the storage unit 3 on the RAM by the CPU. Is realized.

≪Search condition acquisition means≫
The search condition acquisition unit 10 acquires a plurality of predetermined search conditions for a document (search target information) desired by the user as a search condition. In this embodiment, the search condition acquisition unit 10 acquires a plurality of words included in a document and attribute information of the document itself as a search condition.

≪Bitmap information calculation means≫
The bitmap information calculation unit 11 includes an index acquisition unit 13, a decompression unit 14, a bitmap calculation unit 15, and an index acquisition control unit 16.
The index acquisition means 13 displays the bitmap table in order from the largest to the smallest data inclusion relationship based on a plurality of search conditions acquired by the search condition acquisition means 10 from the indexes created in advance. To win. In the present embodiment, the index acquisition unit 13 controls the index k from the index storage unit 6 in the order of the first level (uppermost level), the second level,. (K = 1, 2,...) The bit map table of the stage is acquired.

≪Defrost means≫
The decompressing unit 14 generates a bitmap table in which the compressed data is decompressed when predetermined data stored in the bitmap table acquired by the index acquisition unit 13 is compressed. In the present embodiment, the decompressing unit 14 decompresses data stored in the k-th bitmap table of the index under the control of the index acquisition control unit 16 when the data is compressed.

≪Bitmap calculation means≫
The bitmap calculation means 15 is acquired by the index acquisition means 13, and when the predetermined data is compressed, the compressed data is decompressed. In order from the bitmap table to the smallest one, a plurality of bitmap sequences corresponding to the plurality of search conditions acquired by the search condition acquisition unit 10 are extracted, and the corresponding positions of the extracted plurality of bitmap sequences, respectively. A bit map is calculated by performing a bitwise operation on the bits assigned to. In the present embodiment, the bitmap calculation means 15 is controlled by the index acquisition control means 16 and a plurality of bitmaps extracted from the bitmap table of the k-th index index (after decompression if compressed) to be processed. A bitwise operation is performed on the bitmap sequence. Here, the bitwise operation means to perform a logical product when the input search condition is “AND search”, and when the input search condition is “OR search”, a logical sum is calculated. It means taking.

  The index acquisition control means 16 controls the index acquisition means 13, the decompression means 14, and the bitmap calculation means 15 as a whole. In the present embodiment, the index acquisition control unit 16 determines whether or not the data stored in the k-th bit map table of the index acquired by the index acquisition unit 13 is compressed. First, the data is decompressed by the decompression means 14 and then output to the bitmap calculation means 15. On the other hand, if not, it is output to the bitmap calculation means 15 as it is. Also, the index acquisition control means 16 determines whether or not there is a bitmap table next (below) after the k-th bitmap table of the index acquired by the index acquisition means 13. If there are more bitmap tables, the index acquisition means 13 repeats the same process. On the other hand, if there are no more bitmap tables, the current bitmap The bit map calculated by the calculation unit 15 is output to the bit map correspondence information extraction unit 12 at the subsequent stage.

≪Bitmap compatible information extraction means≫
The bitmap correspondence information extraction unit 12 uses a plurality of documents (hereinafter referred to as bitmap correspondence information) as information corresponding to the bitmap calculated by the bitmap calculation unit 15 for the bitmap table having the smallest data inclusion relationship (hereinafter referred to as bitmap correspondence information). Among the search target information), a plurality of documents corresponding to the bit positions to which any one bit (1 or 0) is assigned in the bitmap calculated for the bitmap table having the smallest data inclusion relationship (search target) Information). In this embodiment, the bitmap correspondence information extraction unit 12 stores information when the index acquisition control unit 16 determines that there is no further bitmap table after (below) the k-th bitmap table. The document corresponding to the bit position to which the bit “1” is assigned in the bitmap calculated by the bitmap calculation means 15 is extracted from the large number of documents stored in the section 7.

<Output means>
The output unit 5 includes a predetermined communication interface and an output interface. The output unit 5 transmits the document (search target information) extracted by the bitmap correspondence information extraction unit 12 to a PC used by the user via a communication network such as the Internet. The output means 5 can also output the document (search target information) extracted by the bitmap correspondence information extraction means 12 to an output device D such as a liquid crystal display connected to the information search device 1 with a cable. .

  Note that the information search apparatus 1 uses a general computer as a search condition acquisition unit 10, a bitmap correspondence information extraction unit 12, an index acquisition unit 13, a decompression unit 14, a bitmap calculation unit 15, and an index acquisition control unit 16. It can also be realized by executing an information search program that functions as: These programs can be distributed via a communication line, or can be written on a recording medium such as a CD-ROM for distribution.

[Operation of information retrieval device]
The operation of the information search apparatus 1 shown in FIG. 1 will be described with reference to FIG. 8 (refer to FIG. 1 as appropriate). FIG. 8 is a flowchart showing the operation of the information search apparatus shown in FIG. First, the information search apparatus 1 acquires a plurality of search conditions related to information desired by the user from the input device M via the input means 2 by the search condition acquisition means 10 (step S11: search condition acquisition step). Then, the information search device 1 initializes the number k of index stages to be acquired by the index acquisition control means 16 and sets k = 1 (step S12). Next, the information search apparatus 1 acquires the k-th bit map table of the index corresponding to the plurality of search conditions acquired in step S11 by the index acquisition means 13 (step S13: index acquisition step). Then, the information retrieval apparatus 1 determines whether or not the data stored in the k-th bitmap table acquired in step S13 is compressed by the index acquisition control means 16 (step S14). If the data has been compressed (step S14: Yes), the information retrieval apparatus 1 decompresses the compressed data by the decompression means 14 (step S15: decompression step). Next, the information search apparatus 1 calculates a bitmap by performing a bitwise operation on a plurality of bitmap sequences extracted from the acquired k-th bitmap table by the bitmap calculator 15 (step S16). : Bitmap calculation step).

  On the other hand, when the data stored in the k-th bit map table is not compressed in step S14 (step S14: No), the information search device 1 skips step S15 and proceeds to step S16. Subsequently, the information search device 1 adds “1” to the value of the number k of the current index (k = k + 1) by the index acquisition control unit 16 to acquire a new k (step S17), and the index It is determined whether or not the k-th bit map table exists (step S18). When the k-th bit map table of the index exists, the process returns to step S13. Here, when the information retrieval apparatus 1 returns, the information retrieval apparatus 1 retrieves the corresponding data in the k-th bit map table of the index from the place where the bit “1” is set in the bit map calculated in step S16. To win.

  On the other hand, if there is no k-th bitmap table in the index in step S18 (step S18: No), the information search device 1 uses the bitmap correspondence information extraction unit 12 to calculate the bit calculated in the immediately preceding step S16. Bitmap correspondence information (document) corresponding to the map (bitmap calculated from the lowest level of the index) is extracted from the information storage unit 7 (step S19: Bitmap correspondence information extraction step). Then, the information search device 1 outputs the bitmap correspondence information (document) extracted in step S19 to the output device D through the output unit 5 as search result information (step S20).

〔Concrete example〕
Next, an outline of the search processing operation of the information search apparatus 1 will be described using a specific example. Here, as an example, an index created in advance based on the bitmap table (modified bitmap table 40) shown in FIG. 4 (the first stage 71 of the index shown in FIG. 7 and the second stage (the lowest stage) of the index). 72) is stored in the index storage unit 6 (see FIG. 1). Further, it is assumed that a large number of documents including each document associated with the modified bitmap table 40 are stored in the information storage unit 7 (see FIG. 1). Furthermore, as shown in FIG. 9, when the search condition input by the user includes two words “excursion” and “fried egg” and searches for a document that is “within the expiration date” (AND search). Is assumed.

  First, the search condition acquisition means 10 of the information search apparatus 1 acquires three search conditions shown in FIG. Subsequently, the index acquisition means 13 obtains “condition group 1” including “excursion” of the three search conditions shown in FIG. 9 from the first level 71 of the index shown in FIG. A bitmap string, a bitmap string of “condition group 2” including “fried eggs”, and a bitmap string of “condition group M” including “expiration date” are acquired. The acquired bitmap sequence is shown in FIG. Here, since the first level 71 of the index shown in FIG. 7, that is, the degenerate bitmap table 50, is not data compressed, the index acquisition control means 16 does not use the decompression means 14. Then, the bitmap calculation unit 15 performs a logical product operation on the bitmap sequence acquired by the index acquisition unit 13, that is, the bitmap sequence shown in FIG. FIG. 11 shows a bitmap (logical product result) calculated at that time.

  In the modified bitmap table 40 shown in FIG. 4, the places where the “1” bits stand are partially concentrated, so that the “1” bit in the degenerate bitmap table 50 (first stage 71 of the index shown in FIG. 7). Since the number of locations where the error occurs is reduced, the cost of bitwise calculation is reduced (effective use of the CPU cache). Furthermore, as shown in FIG. 11, the number of places where the bit “1” stands is reduced even after the bitmap calculation.

  Next, the index acquisition control means 16 determines that it has not yet reached the lowest level of the index after the processing of the first level 71 of the index shown in FIG. Instructs acquisition of the bitmap table. At this time, the index acquisition control means 16 recognizes from the bitmap shown in FIG. 11 that the place where the bit “1” is set corresponds to “Doc group 2”. On the other hand, acquisition of the corresponding data in the bitmap table at the next stage is instructed.

  Next, the index acquisition means 13 acquires the data of the row corresponding to “Doc group 2” in the second level (lowermost level) 72 of the index shown in FIG. An example of the acquired bitmap is shown in FIG. Here, “compressed data 2-1”, “compressed data 2-2”, and “compressed data 2-M” each include a plurality of search conditions. Therefore, the index acquisition control unit 16 uses the decompression unit 14 to decompress the “compressed data 2-1”, “compressed data 2-2”, and “compressed data 2-M”. The decompressed bitmap is shown in FIG. The bitmap calculation means 15 extracts bitmap sequences corresponding to the three search conditions shown in FIG. 9 from the bitmap shown in FIG. 13, and performs a logical product operation. FIG. 14 shows the three bitmap sequences extracted at this time and the calculated bitmap (logical product result).

  Next, the index acquisition control means 16 determines that the lowest level of the index has been reached after the processing of the second level 72 of the index shown in FIG. 7, and the bit map (logical product result) shown in FIG. It outputs to the map corresponding | compatible information extraction means 12. FIG. Then, the bitmap correspondence information extracting means 12 has a Doc id of “6” as a document satisfying all the search conditions from a place where a bit of “1” is set in the bitmap (logical product result) shown in FIG. A document (document title = “Bento Recipe”: see FIG. 4) is acquired from the information storage unit 7. As a result, the information search apparatus 1 outputs the extracted document as search result information.

  According to the present embodiment, the information search apparatus 1 uses an index created in advance from a bitmap table in which a large amount of stored documents (search target information) are associated with all search conditions that can be input. In addition, in the bitmap table that is the basis of the index, the places where the bits of “1” stand are concentrated in groups, so that the compression efficiency is improved and partial decompression is possible, and “1” after degeneration is possible. It is possible to reduce the number of places where the bit stands. Here, there is a high possibility that the place where the bit of “1” stands is a search target desired by the user, and the number corresponds to the number of search candidate groups including the search candidates. When this number is reduced, the number of places where the bit of “1” stands in the bitmap calculation is reduced. As a result, it is possible to handle a plurality of search conditions at once without separating them and processing search resources at high speed while consuming a small amount of resources (CPU cache, memory, disk). Make it possible.

  As mentioned above, although embodiment of this invention was described, this invention is not limited to this, It can implement in the range which does not change the meaning. For example, in this embodiment, the number of index stages is two, but may be three or more. For example, if the reduced bitmap table 50 shown in FIG. 5 is further reduced, a reduced bitmap table 150 can be constructed as shown in FIG. The degenerate bitmap table 150 includes a condition group 151 and a Doc group 152 as items. “Condition group 1” of the condition group 151 is a new group in which “condition group 1” and “condition group 2” in the reduced bitmap table 50 shown in FIG. Further, “Doc group 1” of Doc group 152 is reduced by a superordinate concept (for example, “hobby”) that combines “Doc group 1” and “Doc group 2” of the reduced bitmap table 50 shown in FIG. A new group. That is, the bit “1” indicated by reference numeral 153 corresponds to the bit (1, 0, 1, 1) in the corresponding area of the degenerate bitmap table 50 shown in FIG. By repeatedly reducing or compressing and using a multi-stage index of 3 or more levels, for example, a desired document can be searched quickly from a large number of documents of 10 million or more based on a plurality of search conditions with low consumption resources. can do.

  In the present embodiment, the search conditions and the documents are grouped according to similar ones in the bitmap table based on the index, but the grouping criteria are not limited to this. Further, there are generally many methods for mechanically classifying into subsets or classes (classification, clustering), and the methods are not limited to these methods. It is only necessary to bias the bit arrangement on the bitmap by clustering at least one of the set of search target information and the set of search conditions. In particular, both the characteristics of the bit bias of the “collection of search target information to be stored” created by clustering and the bit bias characteristics of the “set of search conditions” created by clustering are both the same. Are preferably clustered. This can be realized, for example, by performing clustering such that the similar group name on the document side and the similar group name on the search condition side have the same group name.

  Further, in the present embodiment, the index configured in multiple stages by sequentially performing the process of degeneration or compression divides a plurality of documents (search target information) into a plurality of groups based on the similarity of contents, and a plurality of indexes These search conditions are divided into a plurality of groups based on the similarity of contents, but the present invention is not limited to this. For example, the plurality of search conditions in the index may be a plurality of sections configured by dividing predetermined continuous values, and at this time, information obtained by quantifying information related to the plurality of search target information is a plurality of information. It may be included in any of the sections. An example of the bitmap table in this case is shown in FIG. FIG. 16 is a diagram showing an example of a bitmap table using prices as search conditions. Here, as information related to the search target information, a price (integer value) as attribute information of the document is assumed. It is also assumed that there are a plurality of documents ranging from 1 yen to 100,000 yen. A section of 500 yen unit from 1 yen to 100,000 yen is one of the search conditions. According to this, even if it is a search condition for a continuous amount such as numerical information such as a price, a search using a bitmap index is possible.

  As another design modification, in addition to a document (search target information) desired by the user, for example, information related to the document or access to the document is stored in the information storage unit 7 (see FIG. 1). An address (URI: Uniform Resource Identifier) or the like may be stored. In this case, the bitmap correspondence information extraction unit 12 (see FIG. 1), as information corresponding to the bitmap calculated by the logical product operation (bitmap correspondence information), in addition to the document desired by the user, Obtain information related to the document and an address to access the document. As a result, the information search apparatus 1 can also present information related to the document and an address for accessing the document together with the document.

It is a lineblock diagram showing typically the information retrieval device concerning the embodiment of the present invention. It is a flowchart which shows an example of the method of constructing | assembling an index. It is a figure which shows an example of the bitmap table used as the origin which builds an index. It is a figure which shows an example of the bitmap table which modified the bitmap table shown in FIG. 3 by grouping. FIG. 5 is a diagram showing an example of a bitmap table obtained by degenerating the modified bitmap table shown in FIG. 4. FIG. 5 is a diagram showing an example of a bitmap table obtained by compressing the modified bitmap table shown in FIG. 4. FIG. 7 is a diagram illustrating an example of a multi-stage (two-stage) index obtained by stacking the degenerated bitmap table illustrated in FIG. 5 on the compressed bitmap table illustrated in FIG. 6. It is a flowchart which shows operation | movement of the information search device shown in FIG. It is a figure which shows an example of the search conditions containing two words and the information of the document itself. It is a figure which shows an example of the bitmap sequence extracted from the 1st step | level of the index shown in FIG. It is a figure which shows the logical product result using the bitmap sequence shown in FIG. FIG. 8 is a diagram illustrating an example of a bitmap sequence extracted from the second level of the index illustrated in FIG. 7. It is a figure which shows the data which decompress | decompressed the compressed data shown in FIG. FIG. 14 is a diagram illustrating a logical product result using the bitmap sequence illustrated in FIG. 13 and an acquired search condition. FIG. 6 is a diagram illustrating an example of a degenerate bitmap table obtained by further degenerating the degenerate bitmap table illustrated in FIG. 5. It is a figure which shows an example of the bitmap table which used price as search conditions.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Information retrieval apparatus 2 Input means 3 Storage means 4 Control means 5 Output means 6 Index storage part 7 Information storage part 8 Program storage part 10 Search condition acquisition means 11 Bitmap information calculation means 12 Bitmap corresponding | compatible information extraction means 13 Index acquisition means 14 Decompression means 15 Bitmap calculation means 16 Index acquisition control means M Input device D Output device

Claims (10)

An information search device that searches, as search target information desired by the user, search target information corresponding to the input search condition from a plurality of search target information based on a search condition input by the user Because
Search condition acquisition means for acquiring a plurality of predetermined search conditions for search target information desired by the user as the search conditions;
The search condition acquisition means from among indexes constructed by sequentially stacking a plurality of bitmap tables processed in a degenerate or compressed manner by processing a predetermined bit data stored in a bitmap table created in advance. Index acquisition means for acquiring a bitmap table in order from the largest to the smallest inclusive relation of data based on the plurality of search conditions acquired in the step;
Decompression means for generating a bitmap table obtained by decompressing the compressed data when predetermined data stored in the bitmap table obtained by the index obtaining means is compressed;
In the bitmap table acquired by the index acquisition means and decompressed when the predetermined data is compressed, the inclusion relation of the data from the largest to the smallest in order A plurality of bitmap sequences corresponding to a plurality of search conditions acquired by the search condition acquisition means are extracted from the bitmap table and assigned to corresponding positions of the extracted plurality of bitmap sequences. Bitmap calculation means for calculating a bitmap by performing a bitwise operation on the obtained bits;
As the information corresponding to the bitmap calculated for the bitmap table with the smallest data inclusion relation, the one bit in the bitmap calculated for the minimum bitmap table from among the plurality of search target information A bitmap correspondence information extracting means for extracting search target information corresponding to the bit position to which
In the previously created bitmap table, the identification information of the plurality of search target information and the plurality of predetermined search conditions are associated with one of the binary bits,
The information search apparatus is characterized in that the index is hierarchically formed so that predetermined data stored in the bitmap table before processing is included in the bitmap table after processing. The index is
When the plurality of pieces of search target information are divided into a plurality of groups, the group is used as a unit, and predetermined data stored in the bitmap table created in advance is sequentially subjected to a process of reducing or compressing and configured in a multi-stage manner It has been
The information search apparatus according to claim 1, wherein the bitmap table indicating each stage of the index can be joined by a bitwise operation. The index is
A process of reducing or compressing predetermined data stored in the previously created bitmap table in units of the group when the plurality of search target information is divided into a plurality of groups based on the similarity of contents It is a multi-tiered structure that is applied sequentially,
The information search apparatus according to claim 1, wherein the bitmap table indicating each stage of the index can be joined by a bitwise operation. The index is
When the plurality of predetermined search conditions are divided into a plurality of groups, a process for sequentially reducing or compressing predetermined data stored in the bitmap table prepared in advance is performed in units of the groups. It is configured in a shape,
The information search apparatus according to any one of claims 1 to 3, wherein the bitmap table indicating each stage of the index can be joined by a bitwise operation. The index is
When the predetermined plurality of search conditions are divided into a plurality of groups based on the similarity of content, the predetermined data stored in the previously created bitmap table is reduced or compressed in units of the groups. It is configured in a multi-stage shape by sequentially performing processing to
The information search apparatus according to any one of claims 1 to 3, wherein the bitmap table indicating each stage of the index can be joined by a bitwise operation. The plurality of predetermined search conditions in the index are a plurality of sections configured by dividing predetermined continuous values,
Information obtained by digitizing information related to the plurality of search target information is included in any of the plurality of sections,
The index is
2. The multi-stage configuration according to claim 1, wherein a predetermined data stored in the bitmap table prepared in advance is sequentially subjected to a process of degenerating or compressing the section as a unit. Item 6. The information search device according to any one of Items 5 to 6. An information search device that searches, as search target information desired by the user, search target information corresponding to the input search condition from a plurality of search target information based on a search condition input by the user Information search method,
The information search device includes:
A search condition acquisition step of acquiring a plurality of predetermined search conditions for search target information desired by the user as the search conditions;
The search condition acquisition step from among indexes formed by sequentially stacking a plurality of bitmap tables processed in a degenerate or compressed manner by processing a predetermined bit data stored in a bitmap table created in advance. An index acquisition step of acquiring a bitmap table in order from the largest to the smallest inclusive relation of data based on the plurality of search conditions acquired in the step;
A decompression step of generating a bitmap table in which the compressed data is decompressed when the predetermined data stored in the bitmap table obtained in the index obtaining step is compressed;
In the bitmap table acquired in the index acquisition step, and when the predetermined data is compressed, the compressed data is decompressed. In the bitmap table, the inclusion relation of the data is from the largest to the smallest. A plurality of bitmap sequences corresponding to the plurality of search conditions acquired in the search condition acquisition step are extracted from the bitmap table and assigned to the corresponding positions of the extracted plurality of bitmap sequences. A bitmap calculation step for calculating a bitmap by performing a bitwise operation on the obtained bits;
As the information corresponding to the bitmap calculated for the bitmap table with the smallest data inclusion relation, the one bit in the bitmap calculated for the minimum bitmap table from among the plurality of search target information A bitmap correspondence information extraction step for extracting search target information corresponding to the bit position to which
In the previously created bitmap table, the identification information of the plurality of search target information and the plurality of predetermined search conditions are associated with one of the binary bits,
The information search method according to claim 1, wherein the index is hierarchically formed so that predetermined data stored in the bitmap table before processing is included in the bitmap table after processing. The index is
A process of reducing or compressing predetermined data stored in the previously created bitmap table in units of the group when the plurality of search target information is divided into a plurality of groups based on the similarity of contents It is a multi-tiered structure that is applied sequentially,
The information search method according to claim 7, wherein the bitmap table indicating each stage of the index can be joined by a bitwise operation. The index is
When the predetermined plurality of search conditions are divided into a plurality of groups based on the similarity of content, the predetermined data stored in the previously created bitmap table is reduced or compressed in units of the groups. It is configured in a multi-stage shape by sequentially performing processing to
9. The information search method according to claim 7, wherein the bitmap table indicating each level of the index can be joined by a bitwise operation.   An information search program for realizing the function of the information search device according to any one of claims 1 to 6 by a computer.

Images