JP2012230493A - Retrieval device, retrieval method, and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a retrieval device, a retrieval method, and a program which can reduce a data amount held by an index storage unit.SOLUTION: A retrieval device comprises a database, an index storage unit, a reception unit, a calculation unit, a first determination unit, a second determination unit, and an acquisition unit. The database stores a plurality of pieces of data information of which each has a plurality of fields. The index storage unit stores index information associating a partial data string which is a part of a data string registered with a specific field, a hush value of the data string, and position information indicating a position of data information of which the specific field registers the data string. When the first determination unit determines that the partial data string which matches a part of a retrieval data string exists in the index storage unit, and also the second determination unit determines that the hash value corresponding to the partial data string matches a hush value of the retrieval data string, the acquisition unit acquires positional information corresponding to the partial data string.

Description

  The present invention relates to a search device, a search method, and a program.

  2. Description of the Related Art Conventionally, there has been known a search apparatus that includes an index storage unit that stores an index (index) having a data structure such as a B-tree, and uses the index read from the index storage unit to search desired data from a database Yes. The smaller the amount of data held in the index storage unit, the smaller the amount of reading at the time of retrieval, so that the retrieval speed can be increased. For this reason, it is desirable that the amount of data held by the index storage unit is small.

  For example, a key storage area in a leaf node of a B tree in which data (key) registered in each row of a specific field of a table stored in the database and information (value) indicating the registration location of the key are stored Techniques to reduce this are known. In this technology, leaf nodes are divided into two layers, and a key and information (pointer) indicating the position of the second layer data structure corresponding to the key are stored in the first layer data structure. Is done. In the second layer data structure, all values corresponding to the key are stored. For example, when the key is registered in each of the second and fifth rows of a specific field of the table, the value indicating the second row and the value indicating the fifth row are the key. Are stored in the second-layer structure corresponding to. On the other hand, since only one key is stored in the first layer data structure, the key storage area in the leaf node is reduced as a result.

JP 2010-72823 A

  However, in the above-described technique, when there is almost no key with the same value, the index storage unit needs to hold all the keys. Therefore, there is a problem in that the amount of data held by the index storage unit cannot be reduced. is there. The problem to be solved by the present invention is to provide a search device, a search method, and a program capable of reducing the amount of data held by an index storage unit.

  The search device according to the embodiment includes a database, an index storage unit, a reception unit, a calculation unit, a first determination unit, a second determination unit, and an acquisition unit. The database stores a plurality of data information each having a plurality of fields each including a data string. The index storage unit includes a partial data string that is a part of a data string registered in a specific field of the plurality of fields, a hash value of the data string, and data information in which the data string is registered in the specific field. The index information associated with the position information indicating the position in the database is stored. The reception unit receives a search data string. The calculation unit calculates a hash value of the search data string. The first determination unit determines that a partial data string that matches a part of the prosecution data string exists in the index storage unit, and the second determination unit determines a hash value corresponding to the partial data string and a hash of the search data string When it is determined that the values match, the acquisition unit acquires position information corresponding to the partial data string.

  The search method of the embodiment includes a first step of receiving a search data string indicating a data string used for the search, a second step of calculating a hash value of the search data string, and a plurality of data each including a data string A partial data sequence that is a part of the data sequence registered in a specific field of any one of the plurality of data information existing in a database storing a plurality of data information each having a field; The partial data that coincides with a part of the prosecution data string in an index storage unit that stores index information in which a hash value of the data string and position information indicating a position of the data information in the database are associated with each other A third step of determining whether or not a column exists, and the partial data that matches a part of the prosecution data sequence in the third step. A fourth step of determining whether or not the hash value corresponding to the partial data string matches the hash value of the search data string; If it is determined in the fourth step that the hash value corresponding to the partial data string matches the hash value of the search data string, the position information corresponding to the partial data string is acquired. And a step.

  Furthermore, the program according to the embodiment includes a first step of receiving a search data string indicating a data string used for search, a second step of calculating a hash value of the search data string, and a plurality of fields each including a data string A partial data sequence that is a part of the data sequence registered in a specific field of any one of the plurality of data information existing in a database that stores a plurality of data information each having The partial data string that matches a part of the prosecution data string in an index storage unit that stores index information in which a hash value of the data string is associated with position information indicating a position of the data information in the database A third step for determining whether or not there exists, and the portion that matches a part of the prosecution data string in the third step If it is determined that a data string exists in the index storage unit, a fourth step of determining whether the hash value corresponding to the partial data string matches the hash value of the search data string; If it is determined in the fourth step that the hash value corresponding to the partial data string matches the hash value of the search data string, the position information corresponding to the partial data string is acquired. Is a program for causing a computer to execute steps.

The block diagram which shows an example of the search device of 1st Embodiment. The figure which shows an example of a database part. The figure which shows an example of data information. The figure which shows an example of an index memory | storage part. Flow chart showing an example of search processing The flowchart which shows an example of a registration process. 10 is a flowchart illustrating an example of a deletion process. The figure which shows an example of the index memory | storage part of 2nd Embodiment. The block diagram which shows an example of the search device of 2nd Embodiment. The figure which shows an example of a registration process. The figure for demonstrating a modification.

(First embodiment)
FIG. 1 is a block diagram illustrating an example of a schematic configuration of a search device 100 according to the first embodiment. As illustrated in FIG. 1, the search device 100 includes an operation display unit 10, a storage unit 20, and a control unit 30. The operation display unit 10 is a means for displaying various screens and information (for example, search results) regarding the search device 100 and allowing the user to input various operations. Although not shown in detail, the operation display unit 10 includes a display panel that displays various screens and information related to the search device 100 and receives a touch input from a user, and operation devices such as various buttons and a mouse.

  The storage unit 20 includes a database unit 22 and an index storage unit 24. The database unit 22 stores a plurality of data information each having a plurality of fields. As shown in FIG. 2, the database unit 22 stores a plurality of data information DQ. Here, the data information DQ is information indicating a document, and as an example, semi-structured data described in XML (Extensible Markup Language) is adopted as the data information DQ.

  FIG. 3 is a diagram illustrating an example of the data information DQ. As shown in FIG. 3, the data information DQ has a plurality of fields F in which data strings are registered. In the present embodiment, an index is created only for a specific field Fx among a plurality of fields F included in each data information DQ. Detailed contents will be described later. Here, the specific field Fx of each data information DQ is a field in which a character string (data string) constituting an application name (application-title) is registered.

  Returning to FIG. 1, the description will be continued. The index storage unit 24 stores an index (index). The index is a data structure such as a B-tree, for example. In the present embodiment, a B-tree is used as an index, but this is not a limitation, and the type of index is arbitrary. For example, a bitmap index or a function index can be employed. As shown in FIG. 4, the index storage unit 24 includes a root node 25 in the highest layer, a branch node 26 in the intermediate layer, and a leaf node 27 in the lowest layer. The leaf node 27 includes a partial character string that is a part of a character string that constitutes the application name, a hash value of the character string that constitutes the application name, and data information DQ in which the application name is registered in a specific field Fx. The index information associated with the ID (position information) indicating the position in the database unit 22 is stored for each row. The ID (position information) of the present embodiment is information that specifies not only the position of the data information DQ in the database unit 22 but also the position of a specific field Fx in the data information DQ. In the present embodiment, the data length of the partial character string is a fixed length. Here, as an example, the data length is set to 10 characters, but the present invention is not limited to this.

A control unit 30 shown in FIG. 1 is a unit that controls each unit of the search device 100, and includes a control device including a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like. Is done. As a function which control part 30 has,
Receiving unit 31, leaf node specifying unit 32, calculating unit 33, first determining unit 34, second determining unit 35, acquiring unit 36, specifying unit 37, third determining unit 38, fourth determining unit 39, registration unit 40, There are a determination unit 41, a deletion target specifying unit 42, and a deletion unit 43. The above functions (31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43) are executed by the CPU by reading the control program stored in the ROM onto the RAM. Is realized. Not limited to this, at least a part of the above functions (31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43) is realized by a hardware circuit. May be.

  The accepting unit 31 accepts various inputs from the operation display unit 10. For example, the accepting unit 31 can accept an input of a character string (referred to as a search character string) that constitutes an application name used for a search, or can accept an input of data information DQ for registering index information. . It is also possible to accept input of data information DQ to be deleted. The leaf node specifying unit 32 specifies the leaf node 27 corresponding to the input character string. The calculation unit 33 calculates a hash value of the input character string.

  The first determination unit 34 determines whether or not a partial character string that matches a part of the search character string exists in the index storage unit 24 (leaf node 27). When the first determination unit 34 determines that a partial character string that matches a part of the search character string exists in the index storage unit 24, the second determination unit 35 calculates the hash value corresponding to the partial character string and Then, it is determined whether or not the hash value of the search character string matches. When the second determination unit 35 determines that the hash value corresponding to the partial character string matches the hash value of the search character string, the acquisition unit 36 uses the ID corresponding to the partial character string as the index storage unit. 24 (leaf node 27). The acquisition unit 36 can also acquire the data information DQ specified by the ID from the database unit 22.

  The reception unit 31, the leaf node identification unit 32, the calculation unit 33, the first determination unit 34, the second determination unit 35, and the acquisition unit 36 described above can be regarded as a search processing unit that executes a search process described later. . Details of the search process will be described later.

  When the receiving unit 31 receives the input of the data information DQ for registering the index information, the specifying unit 37 includes a character string (a registered character string and a registered character string) that is registered in the specific field Fx of the data information DQ. And the ID (position information) of the data information DQ. The third determination unit 38 determines whether or not a partial character string that matches a part of the registered character string exists in the index storage unit 24 (leaf node 27). If the third determination unit 38 determines that a partial character string that matches a part of the registered character string exists in the index storage unit 24, the fourth determination unit 39 uses the hash value corresponding to the partial character string and Then, it is determined whether or not the hash value of the registered character string matches.

  The registration unit 40 stores index information in which the partial character string of the registered character string, the hash value of the registered character string, and the ID specified by the specifying unit 37 are associated with each other in the index storage unit 24 (leaf node 27). sign up.

  The determination unit 41 corresponds to the partial character string when the third determination unit 38 determines that the partial character string that matches a part of the registered character string does not exist in the index storage unit 24, or the fourth determination unit 39 When it is determined that the hash value to be registered does not match the hash value of the registered character string, the registration location of the index information is determined according to a predetermined rule.

  The accepting unit 31, the leaf node specifying unit 32, the calculating unit 33, the acquiring unit 36, the specifying unit 37, the third determining unit 38, the fourth determining unit 39, the registering unit 40, and the determining unit 41 perform registration processing described later. It can also be regarded as a registration processing unit to be executed. Details of the registration process will be described later.

  The deletion target specifying unit 42 specifies the index information to be deleted among the index information stored in the index storage unit 24. The deletion unit 43 deletes the index information specified by the deletion target specifying unit 42 from the index storage unit 24.

  The reception unit 31, the leaf node specification unit 32, the calculation unit 33, the deletion target specification unit 42, and the deletion unit 43 described above can be regarded as a deletion processing unit that executes a deletion process described later. Details of the deletion process will be described later.

  Next, search processing executed by the control unit 30 will be described. FIG. 5 is a flowchart illustrating an example of the search process. As shown in FIG. 5, first, when the input of the search character string is received by the receiving unit 31 (step S511), the leaf node specifying unit 32 specifies the leaf node 27 corresponding to the search character string (step S512). In the present embodiment, the leaf node identification unit 32 traces the B tree and identifies the corresponding leaf node 27. More specifically, it is as follows. Assume that a character string “database management system” is received as a search character string as shown in FIG. First, in the root node 25, the branch node 26 corresponding to the “database management system” is specified. Here, a branch node 26 is assigned to each line of 50 sounds (A line, Ka line,..., W line), and the first character of the search character string is any of the 50 sound lines. Corresponding branch node 26 is specified depending on whether it belongs to. The “database management system” has a leading character “de” and is considered to belong to “ta line”, and therefore, as a branch node 26 corresponding to “database management system”, a branch node corresponding to “ta line” 26 is identified.

  As shown in FIG. 4, each branch node 26 has a plurality of character strings assigned to the branch node 26 (here, two character strings as an example) of the leaf node 27 corresponding to the character string. A pointer indicating the position is stored. Since the character string of two characters counted from the head of “database management system” is “day”, the leaf node 27 indicated by the pointer “eeee” corresponding to “day” corresponds to “database management system”. It becomes a leaf node 27. As described above, the leaf node 27 corresponding to the “database management system” is specified. The method for specifying the leaf node 27 corresponding to the search character string is not limited to the above example.

  Next, the first determination unit 34 determines whether or not a partial character string that matches a part of the search character string exists in the leaf node 27 (step S513). For example, when the search character string is “database management system”, the partial character strings (“database management system”) of the sixth row index information and the seventh row index information of the leaf node 27 in FIG. Therefore, the result of step S513 is affirmative. When it is determined that a partial character string that matches a part of the search character string exists in the leaf node 27, the calculation unit 33 calculates a hash value of the search character string (step S514). On the other hand, if it is determined that a partial character string that matches a part of the search character string does not exist in the leaf node 27, the search process ends.

  After the above-described step S514, the second determination unit 35 compares the hash value of the search character string with the hash value corresponding to the partial character string that matches a part of the search character string, and the hash value of the search character string It is determined whether there is index information including a hash value that matches (step S515). For example, when the search character string is “database management system”, the hash value (H (database management system)) of the index information in the sixth row of the leaf node 27 in FIG. 4 is the search character string (“database management system”). On the other hand, the hash value of index information (H (storage device having means for acquiring execution information of the database management system)) in the seventh row is the search string (“database management system”). Does not match the hash value.

  When it is determined in step S515 described above that there is index information including a hash value that matches the hash value of the search character string, the acquisition unit 36 acquires the ID of the index information (step S516). Next, the acquisition unit 36 acquires the data information DQ specified by the acquired ID from the database unit 22 and displays it on the operation display unit 10 (step S517). On the other hand, if it is determined in step S515 described above that there is no index information that matches the hash value of the search character string, the search process ends.

  Next, a registration process executed by the control unit 30 will be described. FIG. 6 is a flowchart illustrating an example of the registration process. As shown in FIG. 6, first, when the input of the data information DQ for registering index information is received by the receiving unit 31 (step S611), the specifying unit 37 registers in the specific field Fx of the received data information DQ. A character string (referred to as a registered character string) constituting the application name and an ID (position information) of the data information DQ are specified (step S612). Next, the leaf node specifying unit 32 specifies the leaf node 27 corresponding to the registered character string (step S613). Since this specifying method is the same as the specifying method of the leaf node 27 corresponding to the search character string, detailed description is omitted.

  Next, the third determination unit 38 determines whether or not a partial character string that matches a part of the registered character string exists in the leaf node 27 (step S614). If it is determined in step S614 that the partial character string that matches a part of the registered character string does not exist in the leaf node 27, the determination unit 41 determines the insertion location of the index information of the registered character string according to a predetermined rule. (Step S615), and the process proceeds to Step S620 described later. Assume that the registered character string is “data extraction apparatus, extraction method and program”, and the insertion location is determined in the order of the 50 characters of the registered character string. In the example of FIG. 4, it is determined that the registered character string “data extraction device, extraction method, and program” is inserted between the index information on the third row and the index information on the fourth row. However, the present invention is not limited to this, and the predetermined rule can be set arbitrarily.

  On the other hand, when it is determined in step S614 described above that a partial character string that matches a part of the registered character string exists in the leaf node 27, the calculation unit 33 calculates a hash value of the registered character string (step S616). . Next, the fourth determination unit 39 compares the hash value of the registered character string with the hash value corresponding to the partial character string that matches a part of the registered character string, and the hash value that matches the hash value of the registered character string. It is determined whether there is index information including a value (step S617).

  If it is determined in step S617 described above that there is no index information including a hash value that matches the hash value of the registered character string, the acquisition unit 36 corresponds to the partial character string that matches a part of the registered character string. The data information DQ specified by the ID is acquired from the database unit 22, and the character string constituting the application name registered in the specific field Fx of the data information DQ is acquired (step S618). For example, in the example of FIG. 4, it is assumed that the registered character string is “database management system and program”. In this case, the partial character strings (“database management system”) of the index information on the sixth line and the index information on the seventh line match a part of the registered character string, but the hash values do not match. Therefore, the acquisition unit 36 acquires the data information DQ specified by the ID included in the index information of the sixth row and the index information of the seventh row, and stores the specific field Fx in the data information DQ. Get the character string that composes the registered application name.

  Next, the determination unit 41 determines the insertion position of the index information of the registered character string according to a predetermined rule (step S619), and the process proceeds to step S620 described later. Assume that the registered character string is “database management system and program”, and the insertion location is determined in the order of the 50 characters of the registered character string. In the example of FIG. 4, it is determined that the registered character string “database management system and program” is inserted between the index information on the sixth line and the index information on the seventh line. However, the present invention is not limited to this, and the predetermined rule can be set arbitrarily.

  In step S620, the registration unit 40 associates the partial character string of the registered character string, the hash value of the registered character string, and the ID specified by the specifying unit 37 with the insertion location determined by the determining unit 41. Register index information. For example, when the registered character string is “database management system and program”, “database management system” having a data length of 10 characters from the beginning, a hash value of “database management system and program”, and “database management system and program” "Is associated with the ID of the data information DQ registered in the specific field Fx (the ID specified by the specifying unit 37), and the index information is registered at the insertion location determined by the determining unit 41.

  On the other hand, in step S617 described above, when it is determined that there is index information including a hash value that matches the hash value of the registered character string, the registration unit 40 selects the index of the leaf node 27 (index storage unit 24). The index information in which the partial character string of the registered character string, the hash value of the registered character string, and the ID specified by the specifying unit 37 are registered in the line immediately after the line in which the information is stored or in the immediately preceding line. (Step S620).

  Next, the deletion process executed by the control unit 30 will be described. FIG. 7 is a flowchart illustrating an example of the deletion process. As shown in FIG. 7, first, when the input of the data information DQ to be deleted is received by the receiving unit 31 (step S711), the specifying unit 37 reads the application registered in the specific field Fx of the received data information DQ. A character string constituting the name (referred to as a deleted character string) and an ID (position information) of the data information DQ are specified (step S712). Next, the leaf node specifying unit 32 specifies the leaf node 27 corresponding to the deleted character string (step S713).

  Next, the deletion target specifying unit 42 specifies index information to be deleted among the index information stored in the leaf node 27 (step S714). More specifically, the deletion target specifying unit 42 specifies index information including a hash value that matches the hash value of the deletion character string as index information to be deleted. Next, the deletion unit 43 deletes the index information specified by the deletion target specifying unit 42 from the leaf node 27 (step S715).

  As described above, in the present embodiment, the character string (data string) constituting the application name registered in the specific field Fx of the data information DQ is not registered in the index storage unit 24 as it is. Since the partial character string of the character string constituting the application name, the hash value of the character string, and the ID (position information) are stored in association with each other in the index storage unit 24, the index storage unit 24 holds Data amount to be reduced. Therefore, according to the present embodiment, since the amount of reading at the time of search is reduced, there is an advantageous effect that the search speed can be increased.

(Second Embodiment)
Next, a second embodiment will be described. As shown in FIG. 8, the second embodiment differs from the first embodiment in that the data length of the partial character string stored in the index storage unit 24 (leaf node 27) is variably set. The other parts are the same as those in the first embodiment, and thus the description of the overlapping parts is omitted.

  FIG. 9 is a block diagram illustrating an example of a schematic configuration of the search device 100 according to the second embodiment. As shown in FIG. 9, the control unit 30 is different from the first embodiment in that the setting unit 44 is included in the functions of the control unit 30. The setting unit 44 variably sets the data length of the partial character string to be registered in the index storage unit 24 (leaf node 27) and sets the partial character string to be registered.

  FIG. 10 is a flowchart illustrating an example of a registration process according to the second embodiment. Note that search processing and deletion processing are the same as those in the first embodiment described above, and thus description thereof is omitted here. The contents of steps S911 to S914 in FIG. 10 are the same as the contents of steps S611 to S614 in FIG.

  When it is determined in step S914 in FIG. 10 that a partial character string that matches a part of the registered character string does not exist in the leaf node 27, the determination unit 41 inserts index information of the registered character string according to a predetermined rule. A place is determined (step S915). Since the content of step S915 is the same as the content of step S615 of FIG. 6, detailed description is abbreviate | omitted. After step S915, the setting unit 44 sets the data length of the registered character string to the data length for one character. Then, the first character of the registered character string is set as a partial character string (step S916). For example, when the registered character string is “data extraction device, extraction method, and program”, the first character “de” is set as the partial character string. The method for setting the partial character string is not limited to this, and for example, a data string for two characters from the first character can be set as the partial character string. After step S916, the process proceeds to step S922. Since the content of step S922 is the same as the content of step S620 of FIG. 6, detailed description thereof is omitted.

  If it is determined in step S914 described above that a partial character string that matches a part of the registered character string exists in the leaf node 27, the calculation unit 33 calculates a hash value of the registered character string (step S917). Next, the fourth determination unit 39 compares the hash value of the registered character string with the hash value corresponding to the partial character string that matches a part of the registered character string, and the hash value that matches the hash value of the registered character string. It is determined whether there is index information including a value (step S918).

  When it is determined in step S918 described above that there is no index information including a hash value that matches the hash value of the registered character string, the setting unit 44 extends the data length and sets the partial character string to be registered. (Step S919). More specifically, the setting unit 44 sets the data length of the partial character string to be registered to a value larger than the data length of the partial character string that matches a part of the registered character string.

  For example, in the example of FIG. 8, it is assumed that the registered character string is “database management system and program”. In this case, partial character strings (“data”, “de”, “data database”, “data”) of the index information of the third line, the fifth line, the sixth line, and the seventh line are Although it matches a part of the registered character string, the hash value does not match. Therefore, the setting unit 44 sets the data length of the character string to be registered to a value that is longer by one character than the longest data length among the partial character strings that match a part of the registered character string. In this case, the data length of the partial character string to be registered is 5 characters long, and the setting unit 44 sets “data base” having a data length of 5 characters counted from the first character as the partial character string to be registered. Set. This is merely an example, and the method for setting the partial character string to be registered is arbitrary. In short, any character string having a data length longer than the data length of the partial character string that matches a part of the registered character string may be set as the partial character string to be registered.

  After step S919, the process proceeds to step S920. The contents of steps S920 to S922 are the same as the contents of steps S618 to S620 of FIG.

  If it is determined in step S918 described above that there is no index information including a hash value that matches the hash value of the registered character string, the process proceeds to step S922. Since the content of step S922 of FIG. 9 is the same as the content of step S620 of FIG. 6, detailed description is omitted.

  As described above, in this embodiment, although a partial character string that matches a part of the registered character string exists in the leaf node 27, the index information including the hash value that matches the hash value of the registered character string is included. When it is determined that the leaf node 27 does not exist, the data length of the partial character string to be registered is expanded, but it is determined that the partial character string that matches a part of the registered character string does not exist in the leaf node 27. In this case, the data length of the partial character string to be registered is suppressed (as an example, the data length is suppressed to one character in this embodiment). That is, the capacity of the index storage unit 24 (leaf node 27) can be reduced by reducing the data length of the partial character string to be registered in the unnecessary part while increasing the data length of the partial character string to be registered in the necessary part. There is an advantage.

(Modification)
As mentioned above, although embodiment of this invention was described, each above-mentioned embodiment was shown as an example and is not intending limiting the range of invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

  For example, in each of the above-described embodiments, information (document information) indicating a document has been described as an example of data information stored in the database unit 22, but the present invention is not limited to this, and the type of data information is arbitrary. For example, as shown in FIG. 11, the data information may be a data group G constituting each line of the table data 200. Each data group G (data information) has a plurality of fields F in which data strings are registered. Then, an index as described above can be created for a specific field Fx among the plurality of fields F.

  Further, in each of the above-described embodiments, an index is created only for a specific field Fx of the plurality of fields F included in the data information DQ. An index similar to that in each of the embodiments described above can also be created.

DESCRIPTION OF SYMBOLS 10 Operation display part 20 Storage part 22 Database part 24 Index storage part 25 Root node 26 Branch node 27 Leaf node 30 Control part 31 Reception part 32 Leaf node specific | specification part 33 Calculation part 34 1st determination part 35 2nd determination part 36 Acquisition part 37 identification unit 38 third determination unit 39 fourth determination unit 40 registration unit 41 determination unit 42 deletion target identification unit 43 deletion unit 44 setting unit 100 search device 200 table data

Claims (8)

A database storing a plurality of data information each having a plurality of fields each including a data string;
The partial data string that is a part of the data string registered in the specific field of the plurality of fields, the hash value of the data string, and the data information in which the data string is registered in the specific field An index storage unit for storing index information associated with position information indicating a position in the database;
A receiving unit that receives a search data string indicating the data string used for the search;
A calculation unit for calculating a hash value of the search data sequence;
A first determination unit that determines whether or not the partial data sequence that matches a part of the prosecution data sequence exists in the index storage unit;
When the first determination unit determines that the partial data string that matches a part of the prosecution data string exists in the index storage unit, the hash value corresponding to the partial data string and the search data A second determination unit that determines whether or not the hash value of the column matches;
When the second determination unit determines that the hash value corresponding to the partial data string matches the hash value of the search data string, the position information corresponding to the partial data string is acquired. An acquisition unit;
A search device characterized by that. The receiving unit receives the data information for registering the index information;
When the reception unit receives the data information, a registration data sequence indicating the data sequence registered in the specific field of the data information, a specification unit for specifying the position information of the data information,
A third determination unit that determines whether or not the partial data string that matches a part of the registered data string exists in the index storage unit;
When the third determination unit determines that the partial data sequence that matches a part of the registration data sequence exists in the index storage unit, the hash value corresponding to the partial data sequence and the registration data A fourth determination unit that determines whether or not the hash value of the column matches;
When the third determination unit determines that the partial data sequence that matches a part of the registered data sequence does not exist in the index storage unit, or the fourth determination unit corresponds to the partial data sequence When it is determined that the hash value and the hash value of the registration data string do not match, a determination unit that determines a registration location of the index information according to a predetermined rule;
A registration unit that registers the index information at the registration location determined by the determination unit,
The search device according to claim 1. When the fourth determination unit determines that the hash value corresponding to the partial data string and the hash value of the registered data string do not match, the data length of the partial data string of the registered data string is set. , Further comprising a setting unit that sets a value larger than the data length of the partial data string that matches a part of the registered data string;
The search device according to claim 2. Each of the plurality of data information is information indicating a document.
The search device according to claim 1. The database stores table data in which the plurality of data information is arranged in parallel.
The search device according to claim 1. The position information also specifies the position of the specific field in the data information.
The search device according to claim 1. A first step of accepting a search data string indicating a data string used for the search;
A second step of calculating a hash value of the search data string;
Each of the data strings registered in a specific field of the data information of any of the plurality of data information existing in a database storing a plurality of data information each having a plurality of fields each including a data string In the index storage unit that stores the partial data string that is a part, the hash value of the data string, and the position information that indicates the position of the data information in the database, the index storage unit stores the index data of the prosecution data string A third step of determining whether or not the partial data string that matches a part exists;
If it is determined in the third step that the partial data string that matches a part of the prosecution data string exists in the index storage unit, the hash value corresponding to the partial data string and the search data string A fourth step of determining whether or not the hash value matches;
If it is determined in the fourth step that the hash value corresponding to the partial data string matches the hash value of the search data string, the position information corresponding to the partial data string is acquired. And comprising steps
A search method characterized by that. A first step of accepting a search data string indicating a data string used for the search;
A second step of calculating a hash value of the search data string;
Each of the data strings registered in a specific field of the data information of any of the plurality of data information existing in a database storing a plurality of data information each having a plurality of fields each including a data string In the index storage unit that stores the partial data string that is a part, the hash value of the data string, and the position information that indicates the position of the data information in the database, the index storage unit stores the index data of the prosecution data string A third step of determining whether or not the partial data string that matches a part exists;
If it is determined in the third step that the partial data string that matches a part of the prosecution data string exists in the index storage unit, the hash value corresponding to the partial data string and the search data string A fourth step of determining whether or not the hash value matches;
If it is determined in the fourth step that the hash value corresponding to the partial data string matches the hash value of the search data string, the position information corresponding to the partial data string is acquired. A program for causing a computer to execute steps.

Images